1994 SCVIR Meeting Abstracts

1994 SCVIR Meeting Abstracts

1994 SCVIR Meeting Abstracts The following are abstracts of scientific session papers to be presented at the 19th Annual Scientific Meeting of the Soc...

7MB Sizes 1 Downloads 38 Views

1994 SCVIR Meeting Abstracts The following are abstracts of scientific session papers to be presented at the 19th Annual Scientific Meeting of the Society of Cardiovascular and I~terventional Radiology, March 20-24, 1994, San Diego, California. Authors wIth an * following their name have informed the SCVIR that a potential conflict of interest exists.

Sunday Afternoon • 6C • Scientific Session 1 Venous Stents MODERATOR:

Michael D. Dake, MD, Stanford, CA

1 • 1:30 PM Stents for the Treatment of Venous Obstruction M.D. Dake, MD, Stanford, CA • C.P. Semba, MD PURPOSE: We report the use of intravascular stents for the management of venous occlusive disease in 108 patients. M.ATERIALS AND METHODS: All patients had sigmficant symptoms of venous obstruction referable to central and/ or peripheral lesions due to a variety of underlying conditions. The venous territories treated were the SVC (malignant, n = 29; nonmalignant, '!. = 13), IVC (n = 7), innominate-subclavian (n = 39), IlIac (n = 11), upper extremity (n = 7), and dural sinus (n = 2). 'pretreatment venograms showed a highgrade stenoSIS (> 90% stenosis by diameter) (n = 50) or complete occlusion (n = 58). A variety of stents (mean, 2.9 per patient) were used in this series. RESULTS: Mter stent placement, there was complete (n = 97) or partial (n = 7) resolution of venous obstructive symptoms in 104 (96%). At follow-up (range, 3-3~ months; mean, 12.1 months), symptomatic occlusIOn of the segment in which the stent was placed has occur:ed in 13 cases and significant narrowing o~curred III four. The secondary patency rate remains hIgh at 96%. Complications include one death 14 hours after the procedure; gross hematuria with oral warfarin (n = 1); and extrinsic stent compression (n = 2). There was one case of stent migration, which was successfully corrected percutaneously. CONCLUSION: Our results suggest that venous stents are a safe and effective therapy for a variety of venous obstructions.

2 • 1:45 PM Expandable Stents in the Treatment of Central Obstruction in the Venous System P.C. Lakin, MD, Portland, OR • B.D. Petersen, MD R.E. Barton, MD • B. Uchida, BS* • J. Rosch, MD* R.R. Saxon, MD PURPOSE: The authors evaluated efficacy of expandable Z stents in the treatment of benign and malignant obstruction in the central venous system. MATERIALS AND METHODS: Sixty-one patients with symptomatic central venous obstruction were treated percutaneously with Gianturco-Rosch Z (GRZ) stents. Although 45 of these patients had malignant disease, obstruction was secondary to neoplastic encasement in 40 patients, radiation therapy in three, and benign etiology in 18. Thrombolytic therapy was used prior to stent placement in 17 patients. Follow-up studies consisted of venography or duplex examination and/ or clinical examination. !lESULTS: Stent placement provided initial palliation III all 61 patients. Of 40 patients with malignant obstruction, two were lost to follow-up, primary patency was maintained in 28, five had secondary patency, and five had recurrent symptoms without further treatment. Radiation-induced obstructions have remained patent to the present (21.8 months) or until death (15.5-19.9 months). Of 18 patients with benign obstruction, one was lost to follow-up (1.6 months). In the four deceased patients (2.4-16.6 months), two GRZ stents remained patent and two occluded before death. In the 13 survivors, all stents are patent (3.842.7 months) with five requiring intervention. Five complications (8.2%) related to anticoagulant or thrombolytic therapy occurred with two deaths. CONCLUSION: GRZ stents are effective devices for the treatment of central venous obstruction and have exhibited excellent palliation and long-term patency.


2 • Journal of Vascular and Interventional Radiology January-February 1994

3 • 2:00 PM Venous Stents: Four-year Follow-up G.K. Nazarian, MD, Minneapolis, MN S.A Wegryn, MD • H. Bjarnason, MD D.J. Stackhouse, MD • G.J. Rengel, RN W.R. Castaneda-Zuniga, MD, et al PURPOSE: We describe our updated experience in treating venous stenoses and occlusions with metallic endovascular stents. MATERIALS AND METHODS: Gianturco, Palmaz, and Wallstent stents were placed in 52 patients. Venous stent sites included subclavian (n = 9), innominate (n = 3), sve (n = 4), Ive (n = 4) common iliac (n = 20), external iliac (n = 22), common femoral (n = 10), and superficial femoral (n = 1). Most common indications for stent placement were malignant stenoses and chronic occlusions. Venoplasty and!or urokinase was used as ancillary therapy. Patients received anticoagulants for 3-6 months. Regular follow-up was performed with duplex US. RESULTS: Life-table analysis shows 68% I-year overall and 63% 4-year secondary patency. One-, 2-, and 4-year follow-up is available on 23, 12, and four patients, respectively. Death was a complication of stent placement in two patients. Twelve patients died within 6 months after stent placement of primary disease progression. Although early stent failures were more common in patients with prestent occlusions than with prestent stenoses, I-year secondary patency rates were comparable. Secondary patency rates were only slightly lower in patients with local malignancy than in patients with benign disease. CONCLUSION: Endovascular stent placement provides a nonsurgical alternative for reestablishment of venous flow and symptomatic relief in patients with locally malignant as well as benign vascular disease. 4 • 2:15 PM Use of the Wallstent in the Treatment of Malignant Superior Vena Caval Obstruction

J.F. Dyet, FRCR, Hull, England • AM. Cook AA Nicholson PURPOSE: We assessed the role of the self-expanding Wallstent endovascular prosthesis in the treatment of malignant sve obstruction. MATERIALS AND METHODS: Twenty patients (10 men and 10 women) aged 48-76 years (mean, 63.4 years) had Wallstents implanted percutaneously to relieve the obstruction. There were 15 bronchial carcinomas three metastatic tumors (two breast, one renal), o~e mesothelioma, and one lymphoma. Eighteen patients had undergone radiation therapy and! or chemotherapy, with either failure to remit or later recurrence. Six patients required thrombolysis prior to stent placement.

RESULTS: All patients had rapid relief of obstructive symptoms. This was sustained in 18 to date or at death. Two patients, both of whom underwent lysis before stent placement, developed stent thrombosis following withdrawal of anticoagulation at 4 months and 7 days, respectively. CONCLUSION: The Wallstent performs well in relieving obstruction. Stent placement is usually the only method of relieving obstruction when other treatments fail. 5 • 2:30 PM Stent Placement for Venous Stenoses and Occlusions

M. Saeed, MD, La Jolla, CA • H.J. Knowles, Jr, MD PURPOSE: We investigated the efficacy of balloonexpandable stents in managing venous stenoses. MATERIALS AND METHODS: Stents were used to treat 20 venous lesions (13 stenoses, seven occlusions) in 13 patients. Lesion location was sve (n = 4, all malignant), Ive (n = 2), innominate (n = 4), common iliac (n = 2), external iliac (n = 2), superior mesenteric (n = 2), portal vein (n = 1), splenic vein (n = 1), superficial femoral (n = 1), and a venous bypass graft (n = 1). Thrombolysis preceded stent placement in three patients. RESULTS: The mean pressure gradient was reduced from 13 mm Hg (range, 7-24 mm Hg) to 2 mm Hg (range, 0-6 mm Hg). Symptom relief was obtained in 12 of 13 patients. Mean survival for four patients with malignant sve disease was 3 months (range, 2 days to 5 months). Two lesions in this group recurred at 2 months (successfully redilated) and 5 months. One other lesion (low external iliac) in a patient with multiple prior operations recurred twice over 5 months, requiring a bypass. The remaining patients were clinically improved at a mean of 11 months (range, 4-26 months). CONCLUSION: Palmaz stents are effective in managing venous lesions unsuitable for PTA. 6 • 2:45 PM Hemodialysis Access Salvage with Intravascular Stents

C.A Hackworth, MD, Chicago, IL J.D. Rosenblum, MD • A Liu, MD • J.A Leef, MD PURPOSE: Improved medical therapy and hemodialysis for end-stage renal disease have increased longterm patient survival. However, as patients' lifespans increase, hemodialysis access (HDA) has become more problematic secondary to decreasing number of sites. The purpose of this work was to evaluate intravascular stents as a tool for increasing HDA survival. MATERIALS AND METHODS: Fifteen patients with a total of 16 HDA sites received venous stents (Wall-

1994 SCVIR Meeting Abstracts • 3 Volume 5 Number 1

stent) prior to July 1, 1993. The radiologic files and charts of this group were reviewed through September 1, 1993. All HDA patients were evaluated for thrombosis, new access placement, and additional radiologic intervention after venous stent placement. RESULTS: Five HDA sites required no additional therapy except initial stent placement. Eleven HDA sites thrombosed after stent placement and required repeated thrombectomy. New stenotic foci were detected in 10 of 11 of these cases during or after repeat thrombectomy. Seven of these 11 HDA sites were salvaged with additional stent placement or surgery. Four HDA sites were abandoned and new sites obtained. CONCLUSION: Seventy-five percent ofHDA sites were salvaged with aggressive stent implantation to treat venous stenoses during this study period. This may represent a new important use for intravascular stents. Aggressive stent placement may also significantly increase HDA survival as this population undergoes a longer follow-up.

Sunday Afternoon • 6F • Scientific Session 2 TIPS: Complications and Follow-up MODERATOR:

Janette Durham, MD, Denver, CO

7 • 1:30 PM Complications after TIPS: Predictive Variables A.L. McGee, MD, Dallas, TX • C.R. Rees, MD R.L. Niblett, BS • N.G. Diamond, MD S.P. Lee,MD PURPOSE: The authors analyzed frequency and types of complications after TIPS, and identified predictive variables. MATERIALS AND METHODS: Prospective outcome and complications analysis was performed on 78 patients undergoing TIPS procedures. Indications were variceal bleeding in 69, intractable ascites in eight, and hepatorenal syndrome in one. Thirty-three variables were analyzed for significant relationships by using contingency tables, analysis of variance, or discriminate analysis where appropriate. RESULTS: Major complications were death (n = 12, two due to peritoneal bleeding), peritoneal bleeding without death (n = 3), and other major complications (n = 4; pneumonia, puncture of right hepatic artery, false aneurysm of hepatic artery branch, transient chylous effusion with paralysis of right hemidiaphragm); the remaining 59 patients had no complica-

tions. Best predictors of bleeding were hematocrit, creatinine level, platelet count, ascites, and preprocedure portal vein pressure and portosystemic gradient. Best predictors of death were creatinine and albumin levels, encephalopathy, and ascites. Best predictors of major complications as a group were hematocrit, ascites, portal thrombus, creatinine level, portal vein pressure, and gradient. There was a significant association (P = .0088) between Child-Pugh class and death, with death in 0% of 22 patients in class A, in 12.9% of 31 in class B, and in 32% of 25 patients in class C. CONCLUSION: Child-Pugh class is a good predictor of complications, but some of its components are less important than other variables (eg, creatinine level, hematocrit, platelet count, pressures, portal thrombus) for accuracy of prediction. 8 • 1:45 PM Presentation and Management of TIPS Stenoses and Occlusions M.D. Darcy, MD, St Louis, MO • D. Picus, MD M.E. Hicks, MD • T.M. Vesely, MD D.M. Hovsepian, MD PURPOSE: The goal was to analyze the incidence, presentation, and effectiveness of therapy for stenoses and occlusions developing in TIPS. MATERIALS AND METHODS: Findings in 57 patients were reviewed from our prospectively acquired TIPS database. Patients underwent Doppler US at 1, 2,4, and 6 months. Venography was performed at 6 months or for recurrent bleeding or suspicious US findings. We evaluated the occurrence of stenoses, the mode of presentation, and the technical and physiologic efficacy of repeat stent placement. RESULTS: Sixteen of 57 patients (28.1 %) developed 15 stenoses and three occlusions. Two patients had two separate stenoses. Thirty-three percent were found after recurrent bleeding and 67% at routine screening with US. These occurred an average of 6.5 months after TIPS (range, 3 days to 21 months). Stenoses occurred most commonly in the hepatic vein (69.2%). Repeated stent placement within the existing shunt restored patency in 100%. The mean portosystemic gradient was lowered from 23.5 to 10.0 mm Hg. Ten patients had an average of 6 months follow-up after repeated stent placement. Recurrent stenoses developed in two (20%). CONCLUSION: Stenoses or occlusions occur with moderate frequency. Revision can be accomplished with a high degree of success and provides reasonable intermediate-term patency.

4 • Journal of Vascular and Interventional Radiology January-February 1994

9 • 2:00 PM Midterm. TIPS Patency at Follow-up Portal Venography

R.R. Saxon, MD, Portland, OR • R.E. Barton, MD G. Sahagun, MD • B.D. Petersen, MD F.S. Keller, MD • J. Rosch, MD* PURPOSE: The authors evaluated the durability of successfully created TIPS. MATERIALS AND METHODS: Sixty-three follow-up portal venograms were obtained of 37 patients 3 months to 3 years after TIPS creation. Shunt patency was assessed by measuring the diameter stenosis of each shunt relative to the free end of the metallic stent within the hepatic vein. RESULTS: Initial venograms obtained between 3 and 9 months after shunt placement (mean, 6.0 months) showed luminal narrowing of at least 50% in 18 of 37 patients, while 19 of37 patients had shunts narro,:"ed by less than 50%. Additional venograms were obtamed in 26 patients between 9 and 36 months after TIPS (mean, 12.0 months). Sixteen of these 26 patients had a 50% or greater stenosis of their shunt, while 10 had their shunts narrowed by less than 50%. Among patients with stenotic shunts, the predominant site of luminal narrowing was in the parenchymal tract in 41.3%, the stent site in the hepatic vein in 48.3%, and the free hepatic vein in 10.3%. The portosystemic gradient in patients with stenotic and occluded shunts ( > 50% narrowing) was significantly higher than the gradient in patients with patent shunts (19.1 mm Hg mean vs 10.6 mm Hg mean; P < .01). The overall primary patency was 48.6% at 6 months and 23.1% at 1 year. Aggressive intervention at follow-up venography, including angioplasty and stent placement, resulted in a secondary patency of 96% at 6 months and 1 year. CONCLUSION: Although a patent TIPS can be maintained in almost all patients, midterm follow-up with portal venography reveals a high incidence of shunt stenosis. 10 • 2:15 PM Evaluation of US in Identifying TIPS Complications L.A. Deyoe, MD, St Louis, MO • M.D. Darcy, MD W.D. Middleton, MD • S.A. Teefey, MD J.A. Brink, MD • S.P. Quillin, MD PURPOSE: This study was performed to evaluate the use of US in identifying shunt abnormalities in patients after placement of TIPS. MATERIALS AND METHODS: US was performed on 57 patients at regular intervals after TIPS placement. Velocities were measured in the main portal vein and shunt. Direction of flow was determined in the right and left portal vein. Venography and pres-

sure measurements were performed at 6 months or if US indicated an abnormality. Twenty-three patients had venographic and sonographic correlation and form the patient population for this study. RESULTS: Seven patients had portosystemic gradients of 10 mm Hg or less. Six of seven maintained peak shunt velocities of 100 em/sec or greater (range, 100-189 em/sec; average, 129 em/sec). Five of seven patients maintained retrograde flow in the right and left portal veins. Shunt occlusion was falsely diagnosed at US in one of seven patients. Sixteen patients had elevated portosystemic gradients (range, 14-27 mm Hg). All 16 of these patients had one abnormality or more demonstrated at US. Ten had peak shunt velocity fall below 90 em/sec (range, 0-84 em/sec; average, 47 cm/sec). In 10 patients, the direction of flow changed in the right and/or left portal vein. Three patients had shunt stenosis with a focal area of increased velocity within the shunt (> 200 em/sec). Three had shunt occlusion with no flow identified. Three had hepatic vein narrowing visualized. CONCLUSION: US can reliably monitor TIPS after placement. Peak shunt velocities below 90 em/sec, conversion of flow to antegrade in the right or left portal vein, and focal shunt velocity increases all suggest hepatic vein or shunt stenosis with elevated portosystemic gradients. 11

• 2:30 PM

Duplex and Color Doppler Sonography in the Follow-up Evaluation of TIPS M.C. Foshager, MD, Minneapolis, MN G.K. Nazarian, MD • H. Ferral, MD S.A. Crook, MD • W.R. Castaneda-Zuniga, MD J.G. Letourneau, MD PURPOSE: Our purpose is to assess the accuracy and usefulness of conventional duplex and color Doppler sonography in the follow-up evaluation of TIPS compared with direct venography and to establish expected sonographic findings following uncomplicated TIPS. MATERIALS AND METHODS: Forty patients have been entered into a prospective protocol with routine sonographic and angiographic follow-up. Sonographic evaluation includes a thorough survey of portal venous and hepatic arterial hemodynamics prior to and immediately following TIPS and at 3-month intervals thereafter. Follow-up sonography focuses on stent patency and function. Correlation is made with all concurrent angiographic studies. RESULTS: High-velocity flow (mean peak velocity, 150-200 em/sec) is seen within well-functioning stents. Duplex sonography is 96% sensitive and 100% specific in demonstrating flow within the ste~t and allows accurate prediction of shunt malfunctIOn (stenosis or occlusion) by depicting alterations in shunt

1994 SCVIR Meeting Abstracts • 5 Volume 5 Number 1

velocities. There is a statistically significant increase in hepatic artery peak systolic velocity and main portal vein velocity following TIPS. CONCLUSION: Conventional duplex and color Doppler sonography offer an effective, noninvasive method of evaluating shunt function and portal venous and hepatic arterial hemodynamics following TIPS. Sonography enables accurate prediction of shunt malfunction, thereby facilitating timely therapeutic intervention. 12 • 2:45 PM Portal Venography versus Duplex and Color Flow Doppler US in the Follow-Up of Patients after TIPS v.a. McDermott, FRCR, Durham, NC

P. V. Suhocki, MD • a.E. Newman, MD C.S. Payne, MD • E.R. Hudson, MD T.P. Smith, MD PURPOSE: Early stent thrombosis and gradual development of narrowing within the stent or the hepatic vein are the common causes of TIPS failure. We present a comparison of US and portal venography in the early diagnosis of TIPS failure. MATERIALS AND METHODS: TIPS has been performed on 28 patients over an I8-month period. Follow-up serial US (duplex and color flow) and/or portal venography have been performed every 3-6 months. RESULTS: Twenty-four patients underwent US 1-5 days (mean, 2 days) after TIPS. Four cases of early thrombosis were correctly identified and two treated successfully. Radiologic follow-up longer than 3 months (mean, 8 months) has been obtained on 13 patients, who have undergone 34 US and 18 portal venographic examinations. Both US and portal venography correctly demonstrated all five cases of stenosis in the stent or hepatic vein. No flow in the shunt was seen at US on two occasions. Portal venography confirmed one to be occluded and the other severely stenotic. CONCLUSION: US is a reliable screening test for TIPS occlusion or stenosis. Portal venography is more invasive but offers the advantage of permitting direct measurement of the portosystemic venous pressure gradient and the possibility of treating stenosis immediately following diagnosis.

Sunday Mternoon • 6D • Scientific Session 3 Caval Filters MODERATOR:

Joseph Bonn, MD, Philadelphia, PA

13 • 1:30 PM Retrievable Nitinol Vena Cava Filter: Experimental and Initial Clinical Results N. Nakagawa, MD, Iowa City, IA • AH. Cragg, MD T.P. Smith, MD • F. Castaneda, MD W.H. Barnhart, RTR • S.C. De Jong, BA PURPOSE: We describe the properties of a new retrievable nitinol vena cava filter and report experimental and initial clinical results. MATERIALS AND METHODS: The filter is made of nitinol wire that forms a spiral cone and retrieval wire. The filters were introduced through an indwelling 5.5-F transfemoral sheath into the vena cava of 10 sheep. In seven animals, four 4 x 30-mm radiopaque clots were injected to test thrombus-trapping efficacy. Aspiration thrombectomy was attempted, and the filter was removed. Follow-up venography was performed 1 week after placement in three other animals. After successful preclinical testing, the filter was implanted in two patients. RESULTS: All 10 filters were successfully and easily placed. All filters were thrombus-free at follow-up venography. All clots were trapped by the filter and successfully aspirated. All filters were retrieved without difficulty. Temporary filter implantation was accomplished in two patients for 5 and 7 days. Perisheath thrombosis occurred in both patients. In one patient, infrafilter thrombus was aspirated. CONCLUSION: This study demonstrates the feasibility of temporary vena cava filtration. Potential advantages include easy placement, surveillance, and retrieval. One limitation at present is pericatheter thrombosis, which may be obviated by using a less thrombogenic sheath. 14 •

1:45 PM

Feasibility and Pathologic Effects on the Caval Wall of a New Subcutaneously Tethered Temporary Caval Filter in Swine B.S. Kuszyk, MS, Baltimore, MD A C. Venbrux, MD • J.L. Olson, MD J.H. Anderson, PhD • F.A Osterman, Jr, MD, et al PURPOSE: We evaluated the feasibility of deployment and removal of the "Tempofilter" temporary

6 • Journal of Vascular and Interventional Radiology January-February 1994

caval filter and studied the pathologic effects on the caval wall in swine. MATERIALS AND METHODS: Filters were implanted in the infrarenal IVC of 11 swine via a right external jugular cutdown. Immediate postplacement, mid-study, and pre-explant cavography was performed. Anticoagulation was not used after the procedure. Four swine underwent in situ caval dissections at 1-10 weeks after filter placement. Filters were explanted from the remaining seven animals prior to sacrifice 1-6 weeks after placement. Caval wall samples were evaluated grossly and histopathologically. RESULTS: Filters were successfully deployed in all 11 animals, and were easily explanted in seven of seven attempts up to 6 weeks after placement. Complications included mild cephalic migration, thrombus formation on the tethering catheter with subsequent pulmonary embolism, and caval stenosis. No caval occlusion or perforation was seen. Pathologic analysis showed mild and largely reversible caval wall damage. CONCLUSION: Placement and removal of the Tempofilter up to 6 weeks after placement are feasible. Caval wall damage induced by the filter is mild and largely reversible. Anticoagulation may be needed to minimize thrombus formation with this temporary device.

15 • 2:00 PM Preliminary Clinical Experience with the Gunther Retrievable IVC Filter S.F. Millward, MBChB, Ottawa, Ont E.E. Burbridge, MD • S.J. Markman, MD J. Bormanis, MD • R.A. Peterson, MD PURPOSE: We report our experience with the Gunther retrievable filter. MATERIALS AND METHODS: Seven men and six women (age range, 23-85 years; mean, 58 years) had the temporary filter placed for (1) deep venous thrombosis or pulmonary embolism (PE) and a short-term contraindication to anticoagulation (n = 9); (2) prophylaxis following massive PE (n = 3); or (3) prophylaxis following severe trauma (n = 1). Underlying diseases were benign in 10 patients and malignant in three. Filters were introduced through the right femoral vein in 10, the left in two, and the left internal jugular vein in one. RESULTS: No patient developed recurrent PE with the filter in place. All filters were retrieved without complication 4-14 days (mean, 7 days) after placement. No patient developed recurrent PE between 1 and 12 months after filter removal. Two developed IVC occlusion with the filter in place, but it was removed and a permanent filter was not placed. Two patients with underlying malignant disease required placement of a permanent filter. One patient with a

bleeding disorder developed a massive hematoma at the insertion site and died. There were no other deaths or serious complications. CONCLUSION: The Gunther retrievable filter offers adequate protection against PE and can be safely removed up to 14 days after placement. Patients must be carefully selected for a retrievable filter: A permanent filter may be more appropriate for very unstable patients and patients with underlying malignant disease.

16 • 2:15 PM Intravascular US as an Alternative to Contrast-enhanced Vena Cavography before Filter Placement J. Bonn, MD, Philadelphia, PA • J. Liu, MD L.J. Winer, MD • K.L. Sullivan, MD M.J. Shapiro, MD • G.A. Gardiner, Jr, MD PURPOSE: Intravascular US was analyzed as an alternative to contrast material-enhanced cavography for evaluating the vena cava prior to filter placement. MATERIALS AND METHODS: Rationale for intravascular US included contraindications to iodinated contrast material in six patients with renal insufficiency and three patients with a previous life-threatening reaction to nonionic contrast material, limitations on radiation exposure for one 12-week gravid patient, and one morbidly obese patient who exceeded the weight limits of available fluoroscopy equipment. Intravascular US replaced cavography in the pregnant patient and in the nine patients with contrast material contraindications. In the obese patient, intravascular US was the only imaging modality. RESULTS: In all 11 patients, intravascular US successfully demonstrated iliac vein and vena cava patency, absence of thrombus, location of renal veins, absence of anatomic variants, and caval diameter at the desired filter deployment level. Successful filter placement was confirmed in all 11 patients at intravascular US (n = 1), CT (n = 1) or plain film (n = 3) alone; at intravascular US and another modality (n = 4); and with combinations of other modalities (n = 2). CONCLUSION: Intravascular US is a safe and effective alternative to conventional contrast-enhanced cavography for imaging the IVC prior to filter placement in patients unable to undergo evaluation with iodinated contrast material or ionizing radiation.

1994 SCVIR Meeting Abstracts • 7 Volume 5 Number 1

17 • 2:30 PM Comparison of Titanium Greenfield and Vena-Tech Filters with Use ofthe Jugular and Femoral Approach: Short- and Long-term Follow-up H. Bjarnason, MD, Minneapolis, MN

J. W. Yedlicka, Jr, MD • D.J. Stackhouse, MD D. W. Hunter, MD • K. Amplatz, MD G.K. Nazarian, MD PURPOSE: This prospective study was designed to compare periprocedure complications and short-term and long-term outcome after Vena Tech filter (VTF) and titanium Greenfield filter (TGF) placement. Complications were recorded concerning different entry sites, entry-site thrombosis, filter migration, thrombosis, and filter efficacy. MATERIALS AND METHODS: From March 1990, TGFs were placed by means of the femoral vein approach. From November 1991, VTFs or TGFs were placed, based on random selection, from the jugular approach. All difficulties or complications during placement, as well as entry-site or IVC thrombosis, were recorded. The respiratory movement was recorded on cine, and tilting and position of the filter were evaluated at venography. Follow-up with US of the entry site and IVC and abdominal radiography was performed 3 days, 1 month, 6 months, and then yearly following placement. RESULTS: To date, 95 filters have been placed. Fiftyone patients have died. Eleven deaths were early; three were directly related to massive pulmonary embolism. Both filters were comparable for all evaluated parameters, except asymmetry of the legs, which was more commonly seen with the TGF. CONCLUSION: The two filters appear to have similar properties. Incomplete opening of the VTF from the jugular approach was not seen with the new design. 18 • 2:45 PM Percutaneous Balloon Fenestration of the Aorta to Relieve Ischemic Complications of Aortic Dissection

D.M. Williams, MD, Ann Arbor, MI M.D. Dake, MD • M. V. Marx, MD C.P. Semba, MD • J.C. Andrews, MD G.M. Deeb, MD, et al PURPOSE: Aortic dissections are frequently complicated by obstruction of the mesenteric, renal, or lower extremity arteries. Percutaneous aortic fenestration, by establishing local blood flow between the true and false lumens at a chosen level, potentially has a role in the management of these ischemic complications. We report our experience in 10 cases of percutaneous balloon fenestration of aortic dissections.

MATERIALS AND METHODS: Ten patients with acute (three type A and five type B) or chronic (one type A and one type B) aortic dissections had the following 13 complications: mesenteric ischemia (n = 5), renal failure (n = 2), and lower extremity ischemia (n = 6). Ischemia was documented with pressure measurements in the true and false lumens and angiographic demonstration of arterial stenoses or occlusions. All 10 patients underwent percutaneous balloon fenestration alone of the true lumen. Intravascular US complemented fluoroscopy in four cases. RESULTS: Percutaneous fenestration was technically successful and uncomplicated in all 10 patients. Restoration of flow was successful in three of five patients with mesenteric, one of two with renal, and six of six with lower extremity insufficiency. Two patients with persistent mesenteric and the one patient with renal (n = 1) insufficiency died within 2 weeks. CONCLUSION: Percutaneous fenestration, supplemented when necessary with procedures to expand the true lumen, is moderately successful in relieving infradiaphragmatic ischemic complications of aortic dissection.

Sunday Afternoon • 6E • Scientific Session 4 Dialysis Access: Declotting MODERATOR:

Scott O. Trerotola, MD, Indianapolis,

IN 19 • 1:30 PM Thrombosed Dialysis Access Grafts: Percutaneous Declotting without Urokinase

S.O. Trerotola, MD, Indianapolis, IN G.B. Lund, MD • P.J. Scheel, MD S.J. Sauader, MD • A.C. Venbrux, MD F.A. Osterman, Jr, MD PURPOSE: A technique for percutaneously declotting dialysis access grafts without urokinase by using currently available catheters was evaluated. MATERIALS AND METHODS: Thirty-four clotted grafts were treated in 24 patients. Rather than lysing the clot with urokinase, the clot was macerated and pushed into the central circulation with balloon catheters. The procedure was performed on an outpatient basis. RESULTS: Successful mechanical declotting was performed in all but two patients (92%); however, poor venous outflow caused the procedure to be abandoned in four patients after successful declotting, giving a procedure success rate of 83%. Mean total procedure

8 • Journal of Vascular and Interventional Radiology January February 1994

time was 116 minutes. Eight grafts clotted within 1 week. With use of successful dialysis beyond 1 week as our measure of clinical success, we report a 59% clinical success rate with mean primary patency of 126 days (range, 16-322 days). Two complications occurred: both were emboli to the brachial artery; both were successfully treated with urokinase. No symptomatic pulmonary emboli occurred. CONCLUSION: Mechanical thrombolysis of clotted grafts is possible with currently available catheters, with results similar to those reported with mechanical devices and urokinase. The procedure is relatively inexpensive, safe, and well tolerated. 20 • 1:45 PM Failing Hemodialysis Catheters: Early Experience with Percutaneous Fibrin Sheath Stripping M.R. Crain, MD, Milwaukee, WI M. W. Mewissen, MD • G.A. Fueredi, MD R.A. Beres, MD • R. Paz-Fumagalli, MD R.A. Wertz, MD PURPOSE: Fibrin sheaths around dialysis catheters yield inadequate blood flow rates (BFRs) during hemodialysis. We describe a technical modification of percutaneous fibrin sheath stripping (PFSS) suited for hemodialysis catheters, review its efficacy, and compare its cost with alternative lytic therapy. MATERIALS AND METHODS: From July 1993 to September 1993, 11 patients (10 were outpatients) underwent 12 PFSS procedures to restore functional BFR. After angiographic confirmation of a fibrin sheath, a 0.035-inch wire was advanced through the dialysis catheter and was loop-snared in the IVC with a transfemoral Amplatz snare, allowing multiple overthe-wire "strippings." BFR was followed as a measure of adequate dialysis. Costs of PFSS and 12-hour inpatient fibrinolysis were compared. RESULTS: In all patients, PFSS was angiographically successful without complication, and BFR increased from less than 150 mL/min to greater than 350 mL/min. To date, adequate BFR has been maintained in 10 (mean, 23 days; range, 1-59 days). Repeated PFSS was required in one patient after 52 days of adequate BFR. PFSS costs were $1,840 compared with $13,160 for a 12-hour lytic trial in the intensive care unit. CONCLUSION: PFSS is a safe and effective method for the restoration of BFR in failing hemodialysis catheters complicated by fibrin sheath accumulation. Outpatient performance, repeatability, and low cost make it an attractive alternative to costly inpatient fibrinolysis.

21 • 2:00PM Thrombosed Dialysis Grafts: Percutaneous Balloon Thrombectomy versus Thrombolysis

M.R. Middlebrook, MD, Philadelphia, PA M. C. Soulen, MD • C. Cope, MD R.D. Shlansky-Goldberg, MD • Z.J. Haskal, MD M.J. Pentecost, MD PURPOSE: The utility of mechanically declotting thrombosed dialysis grafts was evaluated. MATERIALS AND METHODS: Forty-five patients had 63 episodes of dialysis graft thrombosis over a 19month period. Thirty grafts underwent thrombolysis with 250,000-1,000,000 U of urokinase. Thirty-three grafts underwent mechanical declotting by placing crossing balloon catheters within the graft. Thrombus from the arterial end was pulled centrally by one balloon and then pushed through the graft into the innominate vein or SVC with the other, from where the thrombus embolized to the lungs. Significant stenoses were dilated. RESULTS: Clinical success, defined as successful dialysis for 1 week after the procedure, was achieved in 88% of grafts treated mechanically and 90% of those treated with thrombolysis. Continuous pulse oximetry showed no change in oxygen saturation in either group, and no clinical signs or symptoms of pulmonary embolism were noted. Average total procedure time was 1.8 hours for mechanical declotting and 3.9 hours for thrombolysis (P < .05). Probability ofpatency (mechanical vs thrombolysis) was 40% versus 30% at 3 months, 30% versus 24% at 6 months, and 13% versus 8% at 12 months (P =NS). CONCLUSION: Mechanical declotting of dialysis grafts is faster, cheaper, and as effective and safe as thrombolysis. 22 • 2:15 PM Pulmonary Embolism Following Hemodialysis Access Thrombolysis/Thrombectomy T.L. Swan, MD, Tucson, AZ • S.H. Smyth, MD* S.J. Ruffenach, MD • S.S. Berman, MD G.D. Pond, MD* PURPOSE: Percutaneous thrombolysis/thrombectomy for occluded hemodialysis access sites is now standard therapy in many regions. After switching from infusion thrombolysis to thrombolysis/thrombectomy, we prospectively evaluated 41 consecutive cases for clinical evidence of pulmonary embolism. We performed ventilation-perfusion lung scanning in a subset of 15 of these patients to detect subclinical embolization. MATERIALS AND METHODS: Acutely occluded polytetrafluoroethylene dialysis grafts were treated with pulsed-spray urokinase followed by balloon catheter thrombectomy. After each procedure, the patient

1994 SCVIR Meeting Abstracts • 9 Volume 5 Number 1

was interviewed about chest pain or dyspnea. Heart rate and O2 saturation were monitored during each procedure. Lung scans were interpreted independently of the authors. RESULTS: Patency was achieved in 39 of 41 cases (95%). No patient reported chest pain or dyspnea. Heart rate and O 2 saturation were unchanged compared with the baseline in all cases. The lung scans revealed perfusion deficits in 14 of 15 patients (93%), primarily small peripheral subsegmental defects. CONCLUSION: This study supports our belief that although pulmonary embolism occurs regularly (and often observably) during percutaneous thrombolysis/ thrombectomy, the perfusion deficit is small and clinically insignificant. We remain concerned about two issues: (1) Thrombectomy without lysis may cause larger perfusion defects, and (2) multiple thrombectomies (with or without lysis) may eventually produce significant morbidity. Further study is needed. 23 • 2:30 PM Initial and Long-term Results of Thrombolysis, PTA, and Stents for Central Vein Obstruction in Patients with an Ipsilateral Dialysis Access Graft T.O. McNamara, MD, Los Angeles, CA S. C. Goodwin, MD PURPOSE: We attempted to establish the safety and long-term results of the use of stents for central vein occlusion or stenosis in patients with an ipsilateral dialysis access graft (DAG) and venous hypertension. Prior reports have suggested a high incidence of restenosis or occlusion if the stents are not dilated to more than 10mm. MATERIALS AND METHODS: Initial and long-term results on nine patients were studied prospectively. Repeated venography and physical examination were performed on all patients at 3-month intervals. The Wallstent was used in eight of nine patients. No stent was dilated to more than 10 mm. The study period was November 1991 to April 1993. RESULTS: One patient developed restenosis and mild recurrent arm swelling after 14 months. Repeated PTA cleared the symptoms and increased the vessel caliber. In one patient reocclusion was associated with migration of one of two adjacent Palmaz stents. The remaining stents have remained patent and asymptomatic. CONCLUSION: The use of Wallstents for central vein obstruction and an ipsilateral DAG is a safe procedure. Intermediate-term results indicate a tendency to restenosis at 12-16 months, which responds well to PTA without additional stent placement. Palmaz stents should overlap to reduce the chance of migration, with resultant restenosis between them.

24 • 2:45 PM Percutaneously Introduced Artificial Venous Valve: Experimental Use in Pigs R. Uflacker, MD, Charleston, SC PURPOSE: We assessed the efficacy and feasibility of a new percutaneous artificial venous valve (PAVV) in an experimental pig model to eventually treat human venous valvular insufficiency. MATERIALS AND METHODS: The PAVV was created by inserting a 0.030-mm polyetherurethane membrane into the lumen (12-14 mm in diameter) of single body Z stents. They were introduced percutaneously through femoral or jugular vein puncture with use of a 10-F sheath under fluoroscopic guidance. The PAVVs were used in pigs, either in the SVC or IVC. RESULTS: In all animals, satisfactory valvular function was achieved at supine or orthostatic cavography. Partial valve thrombosis, behind the valvular membrane, was found in all cases 1 week after insertion, but with patency of the valvular lumen. CONCLUSION: The new PAVVs are functional and able to occlude adequately retrograde venous flow. Valve thrombosis is of concern, and systemic anticoagulation and new, less thrombogenic materials are currently being tested.

Sunday Mternoon • 6C • Scientific Session 5 Angioplasty MODERATOR:

Gary J. Becker, MD, Miami, FL

25 • 3:45 PM Local Delivery of an Antiproliferative Drug with Use of Hydrogel-coated Angioplasty Balloons P.M. Consigny, PhD*, Philadelphia, PA J.J. Barry, PhD* • N.J. Vitali, BS PURPOSE: The purpose of this study was to determine the feasibility of using hydrogel-coated angioplasty balloons as a vehicle to deliver drugs that inhibit the vascular smooth muscle cell (VSMC) proliferation that occurs after angioplasty. MATERIALS AND METHODS: Preliminary in vitro experiments, performed to determine the ability of the tyrphostin RG-50872 to inhibit platelet-derived growth factor (PDGF)-induced VSMC proliferation (3H-thymidine incorporation into DNA), demonstrated that 1 fLM RG-50872 produced almost complete inhibition; this effect was observed with 15 minutes of drug exposure, was reversible, and was not due

10 • Journal of Vascular and Interventional Radiology January-February 1994

to cell toxicity. In further experiments, hydrogelcoated and silicone-coated angioplasty balloons (2.5 mm in diameter x 20 mm in length) were coated with either 10 fLL ofRG-50872 (40 fLM in DMSO) or DMSO vehicle, or were left uncoated. Mterward, each angioplasty balloon was inflated, submerged in 50 mL of culture media, and agitated for 2 minutes to promote drug release. Dilutions of this media were tested for their ability to inhibit PDGF-induced VSMC proliferation in vitro. RESULTS: All hydrogel-coated balloons (n = 5) released sufficient RG-50872 to inhibit proliferation by 95% or more, whereas none of the silicone-coated balloons (n = 4) did (P < .01). DMSO-treated and untreated balloons had no effect on proliferation. CONCLUSION: These findings demonstrate that hydrogel-coated angioplasty balloons can be used for local delivery ofRG-50872 and possibly other antiproliferative drugs. 26 • 4:00 PM Local Adventitial Administration of Heparin via Polymer Matrix Delivery: Effect on Intimal Hyperplasia Following Nitinol Stent Deployment in the Iliac Artery and Vein in the Swine M.S. Johnson, MD, Baltimore, MD AC. Venbrux, MD • M.A. Kobelja, BS P.R. Brown, MD • F.A Osterman, Jr, MD J.H. Anderson, PhD PURPOSE: This study was performed to determine the effect of locally released heparin on the degree of intimal hyperplasia and resultant stenosis following vascular stent placement in the iliac artery and vein in swine, and to evaluate the feasibility of local drug delivery, by means of polymer matrix release, to adventitia of treated vessels in the swine model. MATERIALS AND METHODS: Following iliac arteriography, venography, and intravascular US, iliac artery angioplasty was performed in eight of 10 pigs. Heparin, 100 U /kg, was administered intravenously. Hexagonal-cell cylindrical stents 8 mm in diameter were placed in the external iliac artery and contralateral iliac vein. After stent placement, angiography and intravascular US were performed. Laparotomy was performed through an infraumbilical midline incision. Heparin-laden or blank (control) ethylene-vinyl acetate copolymer matrices were placed adjacent to each vessel with a stent. All animals were followed up with interval angiography and intravascular US until sacrifice, either 1 or 3 months after stent placement. Pathologic specimens were studied grossly and with light microscopy. RESULTS: Preliminary results demonstrate less intimal hyperplasia and resultant stenosis in the heparin-

treated pigs than in the control animals. Further evaluation is in progress. CONCLUSION: Locally administered heparin decreases intimal hyperplasia in the artery and vein of the swine. This method of drug delivery is feasible, safe, and effective. 27 • 4:15 PM Drug Delivery into the Arterial Wall after Angioplasty P.M. Consigny, PhD*, Philadelphia, PA K. T. Miller, MD PURPOSE: One potential approach to the prevention of restenosis after angioplasty is to deliver antiproliferative agents directly to the angioplasty site. The purpose of this study was to determine the time course of drug penetration into the media of the balloon-dilated artery. MATERIALS AND METHODS: Balloon angioplasty of the left and right iliac arteries was performed once for 1 minute in each of four rabbits. A double-balloon catheter was then positioned at the site of angioplasty, and the fluorescent dye PKH26 (MW = 961) was delivered under pressure to simulate drug delivery. Mterward, the arteries were removed, and dye penetration into the media was measured on frozen crosssections with fluorescence microscopy. RESULTS: PKH26 dye perfusion was performed for periods ranging from 5 to 50 minutes at a mean pressure of 189 mm Hg. The depth of dye penetration W, micrometers) into the artery was directly related to dye perfusion time (T, minutes) (D = 0.348 T + 11.958; r = .496; P < .01). This equation predicts complete medial dye penetration in 81 minutes assuming an average intima-media thickness (40.08 fLm). CONCLUSION: This study demonstrates that PKH26 can be delivered to the media of the dilated artery. However, the time required to obtain complete penetration may limit the utility of this double-balloon catheter approach to drug delivery. 28 • 4:30 PM Polymeric Endoluminal Photo-Thermal Paving: Initial Experience in Normal and Dissected Canine Carotid Arteries in Vitro M.J. Slepian, MD*, Tucson, AZ P.K. Campbell, PhD* • K. Berrigan, BS* L. Roth, ME* • C.P. Pathak, PhD* AS. Sawhney, PhD* PURPOSE: Polymeric endoluminal paving is a method for applying thin coatings of polymers to the intima of blood vessels that function as stents, barriers, or drug delivery vehicles. Previously, paving required the use of thermal balloon catheters for poly-

1994 SCVIR Meeting Abstracts • 11 Volume 5 Number 1

mer application. We report on a novel method of paving in which a biodegradable polymer with interspersed photoabsorbent dye may be applied by means of localized photoirradiation. Light absorption of the dye leads to preferential polymer heating. The ability of a catheter-based photothermal paving system to line and support normal and dissected arteries was evaluated in vitro. MATERIALS AND METHODS: Canine carotid arteries (n = 10) were perfused in vitro (saline, 37°C, 90 mm Hg); baseline diameters were determined at intraluminal US (HP-BSC, 30 MHz). In five of 10 vessels US-detectable dissections or flaps were created by means ofpassage of a 0.060-inch steel wire. All arteries were then paved (785 nm; 30 seconds; balloon:artery ratio, 1-1.2:1) with films of poly (E-caprolactone) (PCL)(T ill = 60°C) containing 1% tricarbocyanine and imaging was repeated. The percentage of PCL film in contact with the surface or the completeness of flap sealing was assessed. RESULTS: The polymer layer was readily visualized at US; more than 95% of the applied PCL film layer contacted with the endoluminal surface. All photopaved vessels remained dilated with circular crosssectional profiles. Photopaving successfully tacked dissections with complete flap sealing. CONCLUSION: Catheter-based photoirradiation of biodegradable polymers containing interspersed photoabsorbent dye is an effective method for application of supportive endoluminal paving layers and effective sealing of dissections. Further development of this approach may yield a viable percutaneous clinical method for using biodegradable polymers as endoluminal wall supports and localized drug delivery devices. 29 • 4:45 PM Polyetherurethane Coating Decreases Acute Thrombogenicity of Tantalum Vascular Stents A.B. Fontaine, MD*, Columbus, OH K. Koelling, PhD • D. Spigos, MD G. Christoforidis, MD • J. Clay, MS J. Cearlock, MD, et al PURPOSE: The acute thrombogenicity of coated (polyetherurethane) versus bare tantalum stents was evaluated. MATERIALS AND METHODS: Four stents (two coated with polyetherurethane) were balloon expanded in an 8-mm x 80-cm section of Gore-Tex graft. Eight- and 10-F vascular sheaths were placed percutaneously in the femoral artery and vein of a 75-lb swine under sterile conditions. Platelets were labeled with indium-Ill and reinjected for quantitative assay. The stented graft was connected to the arterial and venous sheaths to create an ex vivo fistula. The fistula was opened for 15 minutes, exposing hepa-

rinized blood to fully expanded stents. The fistula was then closed, flushed until clear, and injected with formalin. The stents were explanted, placed in a radionuclide well counter, and, following this, scanned with light and electron microscopy. RESULTS: Radionuclide counts were slightly greater than 15,000 for coated and 25,000 for bare tantalum stents. Light microscopy demonstrated areas of red thrombus covering multiple areas of the bare stent; coated stents demonstrated focal areas of red thrombus at weld points only. Electron microscopy of coated stents demonstrated an irregular coating with scattered bare areas. Coated areas were devoid of thrombus. Bare areas had some platelets and a few red cells. CONCLUSION: Polyetherurethane coating reduces acute thrombogenicity of tantalum stents. Some focal loss of coating is probably due to shear forces ofballoon dilation. Further investigation is underway. 30 • 5:00 PM Is There a Fundamental Difference in the Injury and Repair Process Following Overdilating Vascular Injury after Placement of Rigid versus Flexible Vascular Stents? A.B. Fontaine, MD*, Columbus, OH • D. Spigos, MD G. Eaton, MD • G. Christoforidis, MD H. Khabiri, MD • S. Jung, MD, et al PURPOSE: The healing response to overdilating injury was evaluated after placement of flexible and rigid vascular stents. MATERIALS AND METHODS: Twelve vascular sheaths were placed bilaterally via femoral arteries in six swine. After angiographic measurement, a total of 12 stents (six flexible and six rigid) were balloon expanded up to 8 mm in diameter in 12 6-mm iliac arteries (30% overdilation). All stents were similar in surface area, gauge, and type of wire (tantalum). The primary difference was six were flexible (low hoop strength [LHSJ) and six were rigid (high hoop strength [HHSJ). Stents were imaged with intravascular US 5 weeks after implantation. The animals were killed and the stented segments removed and examined histologically. RESULTS: HHS stents maintained larger diameters than did LHS stents: the average HHS stent was 7.0 mm (range, 6.2-7.5 mm) versus 6.4 mm (range, 6.07.1 mm) for the LHS stent. In addition, HHS stents developed thicker neointima than did LHS stents: HHS average, 0.96 mm (range, 0.01-2.20 mm) versus LHS average, 0.04 mm (range, 0.01-0.07 mm). CONCLUSION: There is more acute gain in lumen size with rigid (HHS) stents than with flexible (LHS) stents. This is offset by thicker neointimal hyperplastic development (increased late loss oflumen size). Overdilation injury with rigid stents may more closely simulate human PTA restenosis.

12 • Journal of Vascular and Interventional Radiology January-February 1994

Sunday Afternoon • 6F •

Scientific Session 6


Ernest J. Ring, MD, San Francisco, CA

31 • 3:45 PM Adverse Pulmonary Hemodynamic Effects after TIPS: Clinical Correlation M.C. Mountford, MD, Seattle, WA S.J. Althaus, MD • D.M. Coldwell, MD PURPOSE: The pulmonary hemodynamic and clinical effects of the TIPS procedure were evaluated prospectively. MATERIALS AND METHODS: One hundred consecutive patients with cirrhosis, Child class B (25%) or C (70%), were treated with TIPS for active esophagogastric variceal bleeding (96%) or intractable ascites (4%). Pulmonary complications were retrospectively analyzed in all patients. Prospective Swan-Ganz readings of the most recent 25 patients were obtained immediately before and after creation of the shunt. RESULTS: The mean pulmonary pressure increased 5 mm Hgwith a range from 10 mm Hgto -2 mm Hg. The mean IVC pressure increased 3 mm Hg with a range from 10 mm Hg to -2 mm Hg. Overall, 20 patients (20%) developed pulmonary edema. Eight of the patients subsequently developed adult respiratory distress syndrome (ARDS) and two of these patients died of respiratory arrest. Six ofthe 25 patients with SwanGanz measurements developed clinical pulmonary edema and/ or ARDS. CONCLUSION: Significant hemodynamic derangements may occur after TIPS. Increased central venous pressure and increased pulmonary pressure seem to be immediate and direct effects of the TIPS procedure. Careful hemodynamic monitoring may facilitate the clinical outcome. 32


Factors Predicting Early Death of the Child C Patient after TIPS Placement Z.J. Haskal, MD, Philadelphia, PA • C. Cope, MD M.J. Pentecost, MD • M.C. Soulen, MD R.D. Shlansky-Goldberg, MD • R.A. Baum, MD PURPOSE: We identified prognostic factors for poor survival in Child class C patients with variceal hemorrhage who underwent TIPS. MATERIALS AND METHODS: Shunts were placed in 42 consecutive patients with variceal bleeding. Three patients were Child class A, 19 were class B,

and 20 class C. Seventy-six percent of patients were actively bleeding at the time of TIPS. RESULTS: Bleeding stopped in 86% and diminished in 5%. The mean portosystemic gradient of21.4 mm Hg ± 7.2 before TIPS dropped to 10.4 mm Hg ± 4.9 after TIPS. The 30-day mortality was 40%, almost entirely due to multisystem failure; 65% for Child C patients versus 18% for Child A and B patients (P < .01). The mean APACHE II score before TIPS for Child C patients who died within 30 days was 26.7 ± 7.7 compared with 10.1 ± 3.9 for those surviving (P < .0001). Serum liver function indexes, platelet count, white blood cell count, ammonia level, age, and etiology of liver disease did not predict poor survival (P > .05). CONCLUSION: A high APACHE II score before TIPS is highly predictive of early death in the Child class C patient with variceal bleeding referred for TIPS. 33 • 4:15 PM TIPS for the Management of Symptomatic Cirrhotic Hydrothorax L.G. Martin, MD, Atlanta, GA • R.M. Strauss, MD SL. Kaufman, MD • T.B. Boyer, MD PURPOSE: The safety and effectiveness of performing TIPS for the management of symptomatic cirrhotic hydrothorax were investigated. MATERIALS AND METHODS: Over an ll-month period, five patients with Child-Pugh class C cirrhosis requiring frequent thoracentesis (mean, 0.77 per week) for control of dyspnea received a TIPS. RESULTS: No significant procedure or postprocedure morbidity occurred. Mean portosystemic gradient was reduced from 21 to 8 mm Hg. Three occluded shunts were treated at 2,2, and 20 weeks with thrombolysis, angioplasty, and placement of additional stents. Secondary shunt patency was 100% at mean follow-up of 42 weeks. Diuretics were discontinued in three patients and were continued in the two others. Need for thoracentesis was reduced from 0.77 per week to 0.01 per week (P < .001) following TIPS. CONCLUSION: TIPS appears to be a safe and useful technique for the management of patients with symptomatic cirrhotic hydrothorax who require repeat thoracentesis for control of symptoms. Shunt occlusion must be ruled out if respiratory symptoms do not improve or if they recur following TIPS.

1994 SCVIR Meeting Abstracts • 13 Volume 5 Number 1

34 • 4:30 PM Liver Anatomy Applied to TIPS R. Uflacker, MD, Charleston, SC • P. Reichert L.C. D'Albuquerque • A.O. Silva PURPOSE: Liver anatomy necessary to perform TIPS was reviewed. MATERIALS AND METHODS: Acrylic casts of 25 human livers were used to study the vascular and biliary anatomy relevant to TIPS. Size of the casts, d~s­ tance between the hepatic veins and the portal vems, anatomy of the portal bifurcation, and the relationship of the three elements of the portal triad at the level of the portal bifurcation were taken into consideration. RESULTS: The mean measurement of the livers was 24.5 cm (latero-Iateral) and 18.0 cm (anteroposterior). In 64% of the cases there was only one right hepatic vein (RHV), in 4% there were two RHVs, in 24% there was an accessory RHV, and in one case there were three RHVs. The mean distance from the RHVand the portal bifurcation was 4.41 cm (range, 2.7-5.4 cm). In 52% of the cases, the path between the RHV and portal bifurcation was safe and free of major vascular and biliary structures. In 48% of the cases, portal, biliary, and/or arterial structures were seen in the path between the RHV and portal bifurcation. The arteries and bile ducts follow the superior aspect of the portal bifurcation and right portal trunk. CONCLUSION: It is suggested that TIPS puncture be performed in the back and in the lower aspect of the portal bifurcation and right portal trunk. The left portal trunk carries a lower risk of puncture complications. 35 • 4:45 PM Wedge Hepatic and Direct Portal Venous Pressures: Correlation Study during TIPS S. Saddekni, MD, Birmingham, AL M.P. Dohrenwend, MD • M.S. Schwartzberg, MD K.M. Hamrick, MD • R.N. Dimick, MD D.J. van Leeuwen, MD, et al PURPOSE: Wedge hepatic venous pressure (WHVP) has been shown to reflect the pressure in the portal vein, particularly in Laennec cirrhosis. Very few studies, however, were done under the same physiologic conditions or attempted to correlate various etiologies with the findings. We found an excellent opportunity to revisit and reexamine these while performing TIPS. MATERIALS AND METHODS: The study involved 42 patients with portal hypertension who underwent TIPS in a 15-month period. WHVP and then direct portal venous pressure (DPVP) were prospectively recorded. Free, wedge hepatic, and direct portography were performed. The etiology of hepatocellular disease causing portal hypertension was as follows: alcoholic

in 17 alcohol/hepatitis C in five, hepatitis C in five, cryptogenic in nine, autoimmune in three, primary biliary cirrhosis in two, and sclerosing cholangitis in one. RESULTS: WHVP accurately predicted portal hypertension in all 42 patients. WHVP and DPVP were extremely close in the majority of patients (37 patients had < 5 mm difference, 27 patients had a difference of 2 mm or less). In five patients, a difference of up to 7 mm was found, but WHVP was the higher pressure in four of the five. WHVP closely reflected DPVP regardless of the etiologies we encountered. CONCLUSION: WHVP is reconfirmed as a true reflection of the portal vein pressure not only in alcoholic cirrhosis but various other etiologies. This indicates the absence of presinusoidal fibrosis in these conditions. Since WHVP reflects the pressure in the main portal vein, we suggest using the term "cor- . rected portal vein pressure" instead of "corrected SInusoidal pressure." 36 • 5:00 PM Fine-Needle TIPS: Methods and Results S.R. Kerns, MD, Gainesville, FL I.F. Hawkins, Jr, MD • F. W. Sabatelli, MD PURPOSE: The methods and results of 63 consecutive fine-needle TIPS procedures were reviewed retrospectively. . MATERIALS AND METHODS: TIPS were created m 63 consecutive patients with portal hypertension. The most common etiologies were alcoholic cirrhosis, postnecrotic cirrhosis, cryptogenic cirrhosis, and cystic fibrosis; the majority of patients had variceal hemorrhage refractory to sclerotherapy or intractable ascites. A fine-needle coaxial system was used, consisting of a 12-F guiding catheter and a blunt tip Colapinto needle sheathed with an 8-F catheter, through which was placed a sharp 22-gauge puncture needle sheathed with a 5-F catheter. Portal vein puncture was guided by using fluoroscopic landmarks, wedged hepatic venography, or real-time US. RESULTS: Initial technical success was 95% with a primary patency rate of 75% and a secondary patency rate of 87%, over a mean follow-up period of 20 weeks (range, 1-102 weeks). Thirty-day mortality was 6% with no deaths due to the procedure itself. Complications included postoperative fever (n = 4), transient pulmonary edema (n = 1), subendocardial myocardial infarction (n = 1), portal vein thrombosis (n = 2), and portal vein branch occlusion (n = 1). No bleeding complications, arterial injury, or symptomatic hemobilia occurred. CONCLUSION: Fine-needle TIPS is a safe and effective treatment of portal hypertension.

14 • Journal of Vascular and Interventional Radiology January-February 1994

Sunday Afternoon • 6D • Scientific Session 7 MR Angiography MODERATOR:

Gary S. Dorfman, MD, Providence, RI

37 • 3:45 PM MR Angiography as the Primary Imaging Approach to the Patient with Systemic Peripheral Ischemia R.A. Baum, MD, Philadelphia, PA G.A. Holland, MD • J.P. Carpenter, MD M.J. Pentecost, MD • C. Cope, MD PURPOSE: The utility of peripheral vascular MR angiography in planning surgical or angiographic procedures has been demonstrated. This prospective investigation was undertaken to further develop and evaluate an imaging approach that incorporates both MR and conventional angiography in the preoperative assessment of patients with arterial occlusive disease. MATERIALS AND METHODS: Patients referred for imaging because of symptomatic peripheral ischemia were examined first with MR angiography. If an operative plan could be formulated based on the M.R findings, a surgical procedure was performed (WIthout a conventional contrast examination). If better anatomic detail was required or if there was a possibility of percutaneous revascularization, "directed" ar~eri­ ography was performed only in the area of questIOn. Validation ofthe preoperative findings was based on surgical exploration and intraoperative or conventional arteriography. RESULTS: One hundred consecutive patients have been prospectively evaluated. Forty-four patients underwent successful surgical procedures solely on the basis of the MR arteriogram findings. Intraoperative arteriograms or surgical exploration confirmed the preoperative MR findings in all but two cases. Fortyeight patients underwent "directed" angiography, 21 of whom had endovascular procedures. Eight patients with contraindications to MR imaging were referred for conventional diagnostic angiography. CONCLUSION: Peripheral vascular MR angiography can help indicate the appropriate surgical or angiographic intervention. Ifused as the initial i~aging modality in patients with peripheral ische~la, c.onventional diagnostic arteriography may be aVOIded m many cases.

38 • 4:00 PM Femoral MR Angiography versus Conventional Angiography: Preliminary Results

S.F. Quinn, MD, Portland, OR • T.A. Demlon, MD R. W. Hallin, MD • L.R. Eidemiller, MD J. Szumowski, MD PURPOSE: Our interest was to develop a bilateral femoral MR angiography examination that would include the aortic bifurcation to the ankle. MATERIALS AND METHODS: Over an 8-month period, 37 patients underwent conventional angi?graphy and bilateral femoral MR angiograp~ye~amma­ tion. The femoral MR angiography exammatIOns were performed with a 1.5-T unit. We used two-di~en­ sional time-of-flight angiography for all studIes. The anatomic region from the aortic bifurcation to the ankle was divided into four segments. The TR was 33 msec and a TE around 7 msec was selected. The flip angl~ was 60°, and gradient moment nulling was used. An inferior concatenated saturation pulse was used. RESULTS: The femoral MR angiogram could have replaced the conventional angiogram in 56.8% ofpatients (21 of 37). In 43.2% of patients (16 of 37), the femoral MR angiogram could not have replaced the conventional angiogram. The reasons for diagnostic failures with femoral MR angiography included artifact from vascular clips (n = 8) or prosthetic joints (n = 3) missed distal aortic stenosis (n = 1), and suboptimal definition of vessels (usually trifurcation vessels) due to field inhomogeneity (n = 8). Seven of the patients in this group had bypass grafts. CONCLUSION: Femoral MR angiography as we applied it could only replace conventiona.l a~gi~graphy in 56% of patients. Some of the current hmltatIOn~ could be avoided by using supplemental duplex scannm~ . around prostheses and vascular clips. The other hmltations will require advances in MR imaging techniques. 39 • 4:15 PM Accuracy of DSA for Identification of Infrapopliteal Vessels

R. Oser, MD, St Louis, MO • D. Picus, MD M.E. Hicks, MD • M.D. Darcy, MD T.M. Vesely, MD • D.M. Hovsepian, MD PURPOSE: Because conventional film-screen angiography has been reported to inadequate!y visualize infrapopliteal vessels in up to 30% ofpatlents: so~e have advocated use of MR angiography to VIsualIze these vessels. In most of these studies, however, DSA has not been used as a standard part of the examination. The purpose of this prospective study was to evaluate the accuracy ofDSA for visualizing infrapopliteal vessels.

1994 SCVIR Meeting Abstracts • 15 Volume 5 Number 1

MATERIALS AND METHODS: The study population consists of 225 consecutive patients referred for the evaluation of lower extremity ischemia. Standard angiographic technique was employed with the addition of DSA images of the most symptomatic footmost commonly by means of an intraaortic contrast material injection. Visualization of distal vessels was evaluated and compared with intraoperative and/ or postintervention angiograms that served as the gold standard. RESULTS: To date, 180 patients have been enrolled, and 73.8% have required infrainguinal intervention (surgical or endovascular). In 44% of these patients, either intraoperative or postintervention angiograms or findings from surgical exploration were available for comparison. In all but two cases, the initial DSA imaging was equivalent or superior to the follow-up study. In these two patients, both with popliteal artery thrombosis, distal vessels were better visualized after thrombolytic therapy. This may reflect the result of thrombolysis, however, rather than a deficiency of DSA. CONCLUSION: DSA is an accurate and reliable method for evaluating infrapopliteal vessels before either surgery or endovascular intervention. 40 • 4:30 PM Distal Anterior Tibial Artery Pseudo-Lesion: The Ballerina Sign R.A. Baum, MD, Philadelphia, PA G.A. Holland, MD • J.P. Carpenter, MD M. C. Soulen, MD • C. Cope, MD PURPOSE: MR angiography has recently been applied to the peripheral vascular system and has been used to plan bypass surgery in patients with advanced disease. Accurate preoperative imaging is crucial to long-term patency of bypass grafts. The purpose of this investigation was to explain false-positive stenoses identified at MR angiography involving the distal anterior tibial artery. MATERIALS AND METHODS: The distal anterior tibial artery was imaged with two-dimensional timeof-flight MR angiography while the foot was dorsiflexed, plantar-flexed, internally and externally rotated, and in neutral position. Five subjects were examined, including two healthy volunteers and three patients with severe peripheral ischemia. Caliber changes were identified on reconstruction as well as transverse images. Results of intraoperative exploration and angiography were available in the three clinical cases. RESULTS: Changing foot position produced caliber changes in the distal anterior tibial artery where it crosses the extensor hallucis longus muscle and the extensor retinaculum in all five patients. In both healthy volunteers 20%-30% stenoses were produced

when the foot was plantar-flexed. Two of three clinical patients had occlusion of the anterior tibial artery on plantar flexon, while in the third patient the stenosis increased from 30% to 70%. The intraoperative angiogram was identical to the MR angiogram obtained in all three patients with the foot in neutral position. CONCLUSION: When the foot is plantar-flexed, the distal anterior tibial artery is compressed by the extensor hallucis longus muscle and the extensor retinaculum and will produce pseudo-occlusions or stenoses at MR imaging. Imaging the foot in neutral position with either the head or extremity coil restores normal flow and eliminates this artifact. Identification and correction of this anterior tibial pseudolesion has important surgical and angiographic consequences. 41 • 4:45 PM Automated Quantification of Stenotic Lesions from MR Angiograms G.A. Holland, MD, Philadelphia, PA K.B. Gupta, MD • R.A. Baum, MD • C. Cope, MD PURPOSE: The purpose of this study was to determine the accuracy of an automated algorithm for the quantitation of stenoses seen on MR angiographic images. MATERIALS AND METHODS: MR angiography was performed by using two-dimensional time-of-flight technique at 1.5 T in 25 patients who underwent contrast arteriography. MR angiography data from these patients and a phantom were postprocessed by using a recently developed algorithm. The vessels are extracted with a three-dimensional (3D) region-growing technique. With this segmented data, a 3D center line is computed (3DCC) for a given vessel and a stack of oblique images that are locally perpendicular to the center line of the vessel is generated. The following measurements were calculated and plotted from this data set: continuous cross-sectional area, anteroposterior and lateral dimensions, and percentage stenosis perpendicular to the local long axis of the vessel. An interactive graphic system then displays the numerical and image data sets. RESULTS: Vessel extraction was validated in phantom studies to within 2.9%. The 3DCC calculated measurements were validated in phantom and in computer models. Accuracy of area measurements was 98% and reproducibility studies demonstrated a 0.2% deviation. All clinically significant stenoses were identified by radiologists on MR angiographic images and by means of 3DCC analysis of MR angiography data. The 3DCC generally underestimated stenoses by 4%6%. Although there was excellent interreader reliability (kappa = 0.84), readers underestimated stenoses by 12%. This was significantly more than with the 3DCC technique (P < .01).

16 • Journal of Vascular and Interventional Radiology January-February 1994

CONCLUSION: The vascular dimensions and grading of stenoses were more accurate and fully automated. 42 • 5:00 PM Cost Comparison of Peripheral Vascular MR Angiography versus Contrast Angiography in Patients Undergoing Surgical Bypass Procedures

R.A. Baum, MD, Philadelphia, PA • D.P. Yin, PhD G.A. Holland, MD • J.P. Carpenter, MD C. Cope, MD • M.J. Pentecost, MD PURPOSE: In some patients, peripheral vascular MR angiography is being used instead of contrast angiography to plan revascularization procedures. The aim of this study was to compare radiologic charges in patients undergoing peripheral vascular MR angiography to charges generated in patients undergoing conventional contrast angiography. MATERIALS AND METHODS: Radiologic charges were examined in 47 patients who underwent peripheral vascular MR angiography instead of contrast angiography for vascular roadmapping prior to surgical bypass. Radiologic charges were accrued from the time of admission to the time of surgery. These were compared with similar charges generated by a second patient population who underwent traditional contrast angiography prior to surgical bypass. The vascular surgeon, number of patients, and types of bypass procedures in each group were identical. RESULTS: The average radiologic charges were 11% lower in patients who underwent peripheral vascular MR angiography instead of contrast angiography prior to revascularization. CONCLUSION: There was a nominal cost savings in patients who underwent peripheral vascular MR angiography instead of contrast angiography prior to undergoing surgical bypass. These savings may grow as peripheral vascular MR angiography becomes streamlined and takes less time to perform.

Sunday Mternoon • 6E • Scientific Session 8 Gastrostomy MODERATOR:

Roy L. Gordon, MD, San Francisco, CA

43 • 3:45 PM Percutaneous Nutrition in the Pediatric Population

R.B. Towbin, MD, Pittsburgh, PA • R. Kaye, MD T. Bender PURPOSE: In 1979, Sacks and Glotzer introduced the idea of percutaneous placement of a gastrostomy tube. Since then percutaneous gastrostomy (PG) has become an effective alternative to surgery. We present our experience with percutaneous gastrostomy in children to evaluate its safety, effectiveness, and complications. MATERIALS AND METHODS: Since 1986, we have performed 308 percutaneous feeding procedures on children ranging in age from 4 weeks to 22 years, including 154 PGs, 103 percutaneous gastrojejunostomies (PGJs), and 51 button insertions. All PG tubes were placed by means of the antegrade technique. RESULTS: PG tube insertion has been successful in 99% of cases. Failures resulted from restricted access due to the position of the stomach, transverse colon, or liver. Major complications, which included peritonitis, respiratory distress, anterior abdominal wall abscess, and gastric herniation, were identified in less than 4% of cases. No postprocedure shunt infection was identified in children with ventriculoperitoneal (VP) shunts in place prior to PG tube insertion. No transverse colon or liver injury was encountered and no premature PG tube loss occurred. Less than 8% of children with PGJs and gastroesophageal reflux or poor gastric emptying went on to require a Nissen fundoplication. CONCLUSION: PG and PGJ are the procedures of choice for children requiring long-term nutritional support. In a previous report, PG has been shown to be safer and more cost-effective than surgical gastrostomy and Nissen fundoplication. In addition, it is safe to place a PG tube in children with indwelling VP shunts.

1994 SCVIR Meeting Abstracts • 17 Volume 5 Number 1

44 • 4:00 PM Gastrostomy Button Placement through Radiologically Created Tracts: Experience in 26 Patients M.A. Borge, MD, St Louis, MO • T.M. Vesely, MD D. Picus, MD • M.E. Hicks, MD • M.D. Darcy, MD D.M. Hovsepian, MD PURPOSE: We report the feasibility of skin-level (button) gastrostomy placement through radiologically created gastrostomy tracts. MATERIALS AND METHODS: We have placed 26 gastrostomy buttons in 26 children (average age, 75 months; range, 9-235 months). All buttons were placed through gastrostomy tracts created by means of percutaneous fluoroscopically guided gastrostomy. Tract age at button placement averaged 16 weeks. The average tract length measured 3.4 cm (1. 7-6.0 cm). Fourteen Bard mushroom-type buttons and 12 MICKey balloon-type buttons were placed initially. Patients have been followed up for an average of 12 months. RESULTS: Button placement was successful at the initial attempt in 24 of 26 patients (92%). Tract rupture with peritoneal leak occurred in the two initial unsuccessful placements. Both patients underwent successful button placement 1 week later. Minor delayed problems (leak, granulation tissue, valve malfunction, balloon breaking) have occurred in nine patients after an average of 6 months. There were no major complications. CONCLUSION: Button gastrostomy devices are a safe alternative to traditional tube-type gastrostomy devices in the pediatric population. Conversion of radiologic gastrostomy tracts to gastrostomy buttons is feasible and safe. Attention to tract integrity and proper button position is required to avoid complications. 45 • 4:15 PM The Gastrostomy Button: Use in Children

R.B. Towbin, MD, Pittsburgh, PA • R. Kaye, MD C. Fitz • T. Bender PURPOSE: The advantages and disadvantages of the button gastrostomy in the pediatric population are discussed. MATERIALS AND METHODS: Fifty-one gastrostomy buttons have been inserted in children ranging in age from 9 months to 18 years. A variety of button designs are available. In our patient group, 28 buttons were 18-F mushroom tips (Bard) and 23 were 14-F inflatable balloon tips (MIC-Key). In some instances tract dilation was necessary for placement. All buttons were inserted into a mature gastrostomy tract of at least 3 months duration.

RESULTS: No complications resulted from button insertion. Of the 28 mushroom-tip buttons, none have been prematurely lost. Of23 balloon-tip buttons, eight have prematurely ruptured and fallen out. No complications resulted from this unexpected button loss. All were easily replaced. CONCLUSION: Gastrostomy buttons are easy to insert and remain in stable position. The mushroom tips are more difficult to insert, require tract dilation, and occasionally are associated with valve failure and leakage. The balloon-tip buttons are easy to insert, do not require tract dilation, but result in a much higher rate of failure. Both types are well accepted by patients and parents. 46 • 4:30 PM Direct Percutaneous Jejunostomy M.J. Hallisey, MD, Hartford, CT • J.C. Pollard, BA PURPOSE: Enteral feeding has an established role in patients unable to take food orally, but long-term nutritional support with gastrostomy or gastrojejunostomy can be a significant problem in patients with previous surgery, chronic aspiration, abnormal stomach position, or recurrent iatrogenic tube removal. In these patients we have evaluated direct percutaneous jejunostomy (DPJ) tube placement under fluoroscopic guidance. MATERIALS AND METHODS: We attempted 14 DPJ tube placements for enteral feeding in 13 patients by using local anesthesia, fluoroscopic guidance, the Seldinger technique, andjejunopexy. There were nine men and four women with a mean age of 75.2 years (range, 45-95 years). Although percutaneous jejunostomy has been described previously, newer technical modifications and equipment now make DPJ a simple and viable fluoroscopic procedure. RESULTS: DPJ placement was successful in 11 patients. Successful jejunal feeding was achieved in all 11 patients within 24 hours of tube placement. A mean follow-up of 57 days (range, 3-291 days) was obtained in all 11 patients. CONCLUSION: DPJ placement is a viable alternative for the creation and maintenance oflong-term enteral tube feeding, particularly in patients with underlying esophagogastric malignancy, recurrent iatrogenic gastrostomy tube removal, or previous gastric surgery.

18 • Journal of Vascular and Interventional Radiology January-February 1994

47 • 4:45 PM Percutaneous Jejunostomy Replacement in Patients after Esophagectomy R.L. Reichle, MD, Baltimore, MD A.C. Venbrux, MD • R.F. Heitmiller, MD F.A. Osterman, Jr, MD PURPOSE: Patients undergoing esophagectomy often have a surgical jejunostomy tube placed at the time of operation. A small portion of patients who have had these feeding tubes removed require late nutritional support. In this study, we expand our experience on a previously described technique for the percutaneous replacement of a feeding jejunostomy tube, which utilizes markers placed during the initial surgical jejunostomy. MATERIALS AND METHODS: Percutaneous replacement of a feeding jejunostomy tube under fluoroscopic guidance was performed on eight patients. Surgically placed metal clips on a jejunal bowel loop tethered to the abdominal wall were used to guide the initial puncture site for access into the jejunum. Once access was obtained and verified, a feedingjejunostomy tube was placed by using the standard Seldinger technique. RESULTS: Percutaneous replacement of a feeding jejunostomy tube was successful in all eight patients, with one patient requiring two attempts on successive days. There were no immediate complications and, with a mean follow-up of 3.1 months, only one tube has had to be replaced (secondary to leaking). CONCLUSION: Percutaneous replacement of a feeding jejunostomy tube by using surgically placed clips as a guide for access is a safe and effective method for providing late nutritional support in the postesophagectomy patient. 48


Variations in Liver-Colon Anatomic Relationship: Relevance to Interventional Procedures P.F. Jaques, MD· M.A. Mauro, MD, Chapel Hill, NC J. McCall, MD PURPOSE: The authors evaluated the value of CT scans prior to percutaneous interventional procedures involving the liver, biliary tree, or portal venous system. MATERIALS AND METHODS: Abdominal CT scans were reviewed over a 4-month period, and cases were selected that demonstrated variant hepatico-colonic anatomy. Each case was analyzed for the position of the colon, gallbladder, and duodenum in relation to the liver, as well as the morphology of the right and left lobes relating to the portal vein and its branches. RESULTS: Seventeen of 517 abdominal CT scans demonstrated variant hepatico-colonic anatomic rela-

tionships. In seven cases, there was normal liver morphology but the colon was interposed between the chest wall and the liver. Ten cases were characterized by varying degrees of hypoplasia or aplasia of one or more segments of the left lobe. The right portal vein was anteriorly exposed close to the transverse colon and gallbladder. Technical modifications may be needed in the performance of various interventional procedures, including percutaneous biliary drainages, biopsies, and TIPS. CONCLUSION: Variations in liver-colon anatomic relations secondary to hepatic development anomalies may have serious implications in the performance of various types of hepatico-biliary interventional procedures ranging from transhepatic biliary drainage to TIPS.

Monday Mternoon • 6C • Scientific Session 9 Aortoiliac Stents Goetz M. Richter, MD, PhD, Heidelberg, Germany



1:30 PM

Endovascular Grafting of Thoracic Aortic Aneurysms M.D. Dake, MD, Stanford, CA • C.P. Semba, MD D.C. Miller, MD • R.S. Mitchell, MD D.M. Williams, MD • R.P. Liddell, BS PURPOSE: We studied the feasibility, safety, and efficacy of transluminal endovascular grafting for treatment of thoracic aortic aneurysms. MATERIALS AND METHODS: Self-expanding stent/graft devices composed of a stainless steel or nitinol frame covered with a skin of knitted or woven Dacron graft material were implanted in 10 patients with thoracic aneurysms. Four false aneurysms (traumatic [n = 1], contained rupture [n = 2], after coarctation repair [n = 1]) and six true aneurysms were treated. All pathology was distal to the left subclavian artery. All stent/grafts were introduced through a 24-F sheath placed by means of a surgically created femoral arteriotomy (n = 7) or puncture of the infrarenal aorta following retroperitoneal surgical exposure (n = 3). The devices ranged from 28 to 45 mm in diameter and from 5 to 16 cm in length. CT and arteriographic follow-up were obtained in all patients. RESULTS: In all cases, the stent/graft prosthesis was successfully placed without complication; however, in three cases, multiple stents were required to treat the lesion. Mean hospital stay after the procedure was 4.5

1994 SCVIR Meeting Abstracts • 19 Volume 5 Number 1

days. Over a mean follow-up interval of 5.8 months (range, 3-15 months) all stentl grafts are patent with persistent thrombosis of the excluded aneurysm. Three cases of transient pleuritic chest pain developed after placement. CONCLUSION: Our initial experience suggests transluminal endovascular grafting offers a promising treatment alternative to standard resection surgery in selected thoracic aortic aneurysm cases. Further investigation is necessary to establish its long-term safety and efficacy. 50

1:45 PM

Aneurysm and Pseudoaneurysm Balloon-expandable Stent/Graft Bypass: Clinical Experience F.J. Rivera, MD, San Antonio, TX J.C. Palmaz, MD* • C.E. Encarnacion, MD T.O. McNamara, MD • M. Saeed, MD • M. Howard, et al PURPOSE: Stents function as architectural support for blood vessels but also for prosthetic graft material. We report on a balloon-expandable stent/graft prosthesis introduced through a 10-12-F sheath and expanded to 8-12 mm in diameter. MATERIALS AND METHODS: Seven patients with aneurysm, arteriovenous (AV) fistula, or pseudoaneurysm were treated. Palmaz balloon-expandable PS30 stents were used in combination with polytetrafluoroethylene grafts (3 or 4 mm). The stentlgrafts were expanded by using conventional PTA balloons. RESULTS: All procedures resulted in technical success. AV fistulas closed completely, and the aneurysms and pseudoaneurysm were completely excluded. All patients have undergone a normal clinical examination. CONCLUSION: A balloon-expandable stentl graft prosthesis shows great promise with high technical success and good preliminary results. 51


Iliac Artery Stent Placement with the Palmaz Stent: Follow-up Study K.D. Murphy, MD, San Antonio, TX J.C. Palmaz, MD* • C.E. Encarnacion, MD VA Le, MD PURPOSE: We evaluated the long-term efficacy of iliac artery stent placement with the Palmaz stent for treatment of limb ischemia. MATERIALS AND METHODS: Iliac stent implantation for limb ischemia was performed in 108 patients. Eighty-three patients (76.9%) were followed up clinically from 1 to 70 months (mean, 25.8 months), and 36 patients (33.3%) were followed up with angiography from 1 to 48 months (mean, 10.4 months). Pa-

tients were classified by using ischemic ranking profile into six categories. Clinical success was defined as a minimum one-stage improvement from preprocedure ranking. RESULTS: Clinical success rate was 98.9% immediately after the procedure, 89.3% at 12 months, and 86.2% at 48 months. Success was statistically more common in patients with higher preprocedure and lower immediate postprocedure ischemic ranking, with unilateral stent placement, and among nondiabetics. Success versus number of stents, stent-Iesion ratio, or postprocedure smoking status was not statistically significant. Follow-up arteriography demonstrated a primary patency rate of 88.9%, four occlusions (11.1%), and only two patients (5.5%) with stenosis of 50% or more. Average stenosis was 17.8% at 6 months and 24.7% at 2 years. Stent stenosis was statistically more common with a higher postprocedure pressure gradient. Thirty-day mortality was 0.9%; overall4-year mortality was 13.8%. CONCLUSION: Long-term clinical and angiographic follow-up demonstrates iliac artery stent placement with the Palmaz stent is efficacious for treatment of limb ischemia. 52

2:15 PM

Self-expandable Stents in the Treatment of Complete Iliac Occlusion: Midterm Follow-up R. Reyes, MD, Canary Islands, Spain J.M. Pulido-Duque, MD • E. Gorriz, MD V. Cabrera, MD • P. Rubio, MD • M. Maynar, MD PURPOSE: We report the use of endovascular stents in the treatment of 58 iliac occlusions. MATERIALS AND METHODS: Self-expanding Wallstents were used to treat 58 iliac occlusions in 56 patients: 52 men, four women; age range, 32-86 years (mean, 62 years). Occlusion length ranged from 3 to 25 cm (mean, 10.3 cm). The number of stents placed in the patients was as follows: one stent (n = 23), two stents (n = 16), three stents (n = 11), and four stents (n = 7). In 30 patients an ipsilateral femoral puncture was used, in nine a contralateral approach, and in 18 a through-and-through technique. The stents were placed immediately after recanalization. Follow-up ranged from 1 to 37 months (mean, 17). RESULTS: Vascular restoration was achieved in all patients but one. Immediate complications were distal embolization (n = 2) and acute stent thrombosis (n = 3); all five conditions were resolved with urokinase infusion. Complications during follow-up were occlusions (n = 4). In two of the patients, the occlusion was cured with urokinase, and two patients underwent bypass surgery. Two patients were lost to follow-up. Follow-up showed a primary patency rate of 82.1% (46 of 56) and secondary patency rate of 94.6% (53 of 56).

20 • Journal of Vascular and Interventional Radiology January-February 1994

CONCLUSION: Endovascular stents offer high technical success in recanalizing iliac occlusions, and their use is an alternative treatment to surgical bypass in these situations. 53 • 2:30 PM Percutaneous Revascularization of Chronically Occluded Iliac Arteries with Use of Thrombolysis and Wallstents T.P. Murphy, MD, Providence, RI G.S. Dorfman, MD • R.E. Lambiase, MD WI Carney, Jr, MD • C.J. Morin, MD C.J. Ashworth, MD PURPOSE: The authors assessed the effectiveness of percutaneous revascularization of chronically occluded iliac arteries with thrombolysis and Wallstents. MATERIALS AND METHODS: Twenty-six chronically occluded iliac arteries (10 common, 15 external, and one both) were recanalized and thrombolysed, and stents were implanted. No patient was refused percutaneous therapy over a I-year period. Twelve patients had severe claudication, nine had rest pain, three had ischemic tissue loss, and two had gangrene. RESULTS: Technical success was achieved in 92%; one failure was due to acute thrombosis due to common femoral artery disease, and another was due to iliac artery rupture. Major complications were seen in two patients (8%). The mean postprocedure transiliac gradient was 1 mm Hg ± 2. Ankle-brachial indexes improved from mean 0.38 ± 0.20 to 0.76 ± 0.21. Average length was 9 cm ± 2. Follow-up to 14 months (mean, 5.5 months ± 3.3) demonstrated 17 asymptomatic patients, five with mild claudication, and one with severe claudication. Primary patency is 88%, primary assisted patency 96%, and secondary patency 100%, excluding technical failures. No mortality occurred within 30 days. CONCLUSION: Percutaneous revascularization of chronically occluded iliac arteries has demonstrated success regardless of severity of symptoms or length of the involved segment and is the initial treatment of choice for all patients. 54 • 2:45 PM Aorto-Iliac Stent Placement versus Surgery: Comparison of Outcomes and Costs J.F. Dyet, FRCR, Hull, England D.M. Steventon, MB, BS • A. Mavor, FRCS PURPOSE: We compared and contrasted the benefits and costs of aorto-iliac stent placement versus surgery in comparable patient groups. MATERIALS AND METHODS: Between 1990 and 1992, 37 patients had stents implanted for aorto-iliac occlusive disease. In 1988 and 1989 (before stent

placement was undertaken in this institute), 116 patients underwent aorto-femoral surgery for aorto-iliac occlusive disease; of these, 34 would today be considered suitable candidates for stent placement. These two groups of patients were then compared for immediate outcome, outcome at 1 year, and cost ofprocedure. RESULTS: Thirty-one patients in the stent group underwent an uncomplicated procedure, five had minor complications (hematomas), and there was one stent thrombosis at 12 hours. At 1 year, 31 maintained improvement, there were three unrelated deaths, and three patients had further procedures. The average cost (including additional procedures) was £2,172 per patient. Eighteen patients in the surgical group patients underwent an uncomplicated procedure. There were eight minor complications; seven necessitated a further operation, and there was one perioperative death. At 1 year, 24 maintained improvement, six were unchanged, and three required a further operation. The average cost was £6,469 per patient. CONCLUSION: In suitable patients, stent implantation offers significant advantages over surgery in both outcomes and cost.

Monday Afternoon • 6F • Scientific Session 10 TIPS: Clinical Results MODERATOR:

Michael D. Darcy, MD, 8t Louis, MO

55 • 1:30PM Five-year Results of TIPS: Technical Standard and Long-term Clinical Efficacy G.M. Richter, MD, Heidelberg, Germany G. Noeldge, MD, • T.K. Roeren, MD G. W. Kauffmann, MD PURPOSE: In 1988 we clinically introduced TIPS and, during the last 5 years, successfully performed the procedure in 137 of 149 patients. This experience has helped to simplify and standardize the technique, which is based on the use of the Palmaz iliac stent. MATERIALS AND METHODS: In the first 18 patients, we tried a simultaneous transjugular and percutaneous transhepatic approach. From patients 17 on, we successfully used combined sonographic and fluoroscopic puncture guidance. Also, we designed a new puncture needle to minimize the risk and optimize the performance. RESULTS: The average portosystemic gradient dropped from 29 mmHg ± 5 to 12 mm Hg ± 4. The

1994 SCVIR Meeting Abstracts • 21 Volume 5 Number 1

30-day encephalopathy rate (de novo) was 7.2%, the rebleeding rate was 6.5%, and the mortality rate was 6.3%. During follow-up (0-60 months; average, 21 months), the 6-month survival rate was 91%, the I-year survival rate was 74%, and the 3-year survival rate was 42%. Shunt restenosis proved to be a significant problem; reintervention was needed at 3 months after TIPS in 46% of all patients, 6 months after TIPS in 17%, and 12 months after TIPS in 6%. The reintervention was strongly associated with the coagulation status and the liver function of the patient: the better the liver function, the higher the restenosis rate. CONCLUSION: The concept of TIPS is here to stay with proved clinical efficacy and low procedure morbidity. A variety of technical aspects have been clarified during our long-term clinical application. Factors such as type of stent, optimal shunt diameter, relative reduction of the portosystemic pressure gradient, and, furthermore, adjunctive medication and the significance of postprocedure care deserve further discussion. 56 • 1:45 PM TIPS in Variceal Bleeding C.E. Encarnacion, MD, San Antonio, TX O.A. Alvarez, MD • K.D. Murphy, MD V.A. Le, MD • J.C. Palmaz, MD* • F.J. Rivera, MD PURPOSE: The role of TIPS in the treatment ofvariceal bleeding was determined. MATERIALS AND METHODS: TIPS was performed in 62 patients for variceal bleeding. Mean age was 49 years (range, 29-84 years). Child-Pugh status was class A in seven, Bin 27, C in 28. Average transfusion requirement before the procedure was 17 units of blood products. RESULTS: Mean follow-up time was 6 months (range, 0-25 months). The rate of variceal rebleeding was 25% (15 of60 patients). Mean transfusion requirement was 2.5 units 30 days after the procedure. Of 20 portograms obtained 6 months after TIPS 11 shunts required revision, and nine did not. Prim~ry patency was 58%, primary assisted patency was 100%, and secondary patency was 98%. Thirty-day mortality reached 25% and overall mortality was 41%. Mean serum bilirubin level 2 weeks after TIPS was 14 mg/dL in patients who died and 3.9 mg/dL in survivors. CONCLUSION: TIPS was effective in arresting variceal bleeding and reducing transfusion requirements. Close follow-up and revisions are the key to long-term patency. High serum bilirubin levels 2 weeks after the procedure was associated with high mortality rate (P = .0008).

57 • 2:00 PM TIPS: Clinical Results R.E. Barton, MD, Portland, OR • R.R. Saxon, MD G. Sahagun, MD • J. Rosch, MD • F.S. Keller, MD PURPOSE: The authors assessed the clinical effectiveness of TIPS and the incidence of complications in patients undergoing TIPS. MATERIALS AND METHODS: One hundred eigh~ teen patients undergoing TIPS for variceal bleeding (n = 106) and ascites (n = 12) were followed up for periods ranging from 3 months to 3 years. Effectiveness of TIPS in relieving presenting symptoms and the incidence of early and late complications were assessed. RESULTS: Ascites resolved in all patients treated for that indication. Bleeding recurred within 30 days in 15 of 106 patients treated for variceal hemorrhage. Five of these had occluded or stenotic shunts. In the other 10, shunts were patent. The 30-day mortality rate was 8.5% (n = 10). Only one death was due to a technical problem occurring during the procedure (hepatic artery injury). In a group of patients followed at least 1 year (n = 60), the 30-day mortality was 13% (n = 8), and the I-year mortality was 25% (n = 15). Hepatic failure was the most common cause of death (n = 6). Death rates were significantly higher in Child class C patients than in others. The incidence of rebleeding in this group was 10% in the first 30 days and 18% overall. Late rebleeding was always associated with an occluded or stenotic shunt, whereas early rebleeding commonly occurred with a patent shunt. Rates of encephalopathy were difficult to determine, but severe encephalopathy occurred only in patients dying of liver failure. No patient surviving 1 year had incapacitating encephalopathy. CONCLUSION: TIPS can be performed safely in virtually all patients. It is an effective therapy for variceal hemorrhage and ascites. Further study is required to better assess the incidence of late complications, such as hepatic failure, and to evaluate long-term shunt patency. 58 • 2:15 PM Doppler Sonographic Evaluation of TIPS Patency and Stenosis M.D. Patel, MD, San Francisco, CA V.A. Feldstein, MD • J.M. LaBerge, MD PURPOSE: The accuracy of Doppler sonography in determining patency and indicating stenoses of TIPS was evaluated. MATERIALS AND METHODS: One hundred sixtyf?ur Doppler US examinations, performed on 61 patIents after TIPS, were retrospectively evaluated for demonstration of shunt patency, maximum peak shunt velocity, and direction of intraparenchymal por-

22 • Journal of Vascular and Interventional Radiology January-February 1994

tal venous (PV) flow. Eighty-one transjugular portal venograms obtained of these 61 patients were independently reviewed. RESULTS: Doppler US documented TIPS occlusion in 20 of 21 (95%) and confirmed shunt patency in 143 of 143 (100%), yielding positive (PPV) and negative (NPV) predictive values of 1.0 and 0.99, respectively. Peak velocity within the TIPS of 0.5 m/sec or less correlated with greater than 75% diameter stenosis at portal venography in 12 of 20 (60%) and was not present in 46 of 47 (98%) without significant stenosis (PPV, 0.92; NPV, 0.85). Interval change in intraparenchymal PV flow direction from hepatofugal to hepatopetal indicated TIPS stenoses in seven of seven (100%) and was not found in 23 of29 (79%) without stenosis (PPV, 0.54; NPV, 1.0). CONCLUSION: Doppler sonography is highly accurate in determining TIPS patency. Determination of peak TIPS velocity and intraparenchymal PV flow direction is useful in evaluating TIPS for stenosis. 59 • 2:30 PM Influence of TIPS on Operative Time and Intraoperative Blood Product Use in Liver Transplant Surgery J.C. McDermott, MD· M.M. Wojtowycz, MD, Madison, WI • M.R. Schuster, MD I.A. Sproat, MD • A.B. Crummy, MD M. Kalayoglu, MD, et al PURPOSE: The effects of TIPS on operative time and intraoperative blood transfusions during liver transplantation were determined. MATERIALS AND METHODS: In 1992, 11 patients underwent liver transplantation at our institution 2-256 days (mean, 66 days) after TIPS placement. In this retrospective study 11 controls were selected (without knowledge of operative details) from those undergoing transplantation over the same period of time without prior TIPS. Controls were selected to match for Child-Pugh classification, history of prior variceal hemorrhage, underlying liver disease, and history of previous abdominal surgery. RESULTS: Operative time was not affected by TIPS (mean, 9.8 hours vs 9.6 hours for controls), and no significant differences were found in the mean numbers of units of platelets (37 vs 41 for controls) and fresh frozen plasma (54 vs 53 for controls) transfused. The mean number of packed red blood cell units given was lower in the TIPS patients (22.6 vs 29.7 for controls) and this difference persisted when cellsaver use was added (25.2 vs 31.5 for controls). Nevertheless, these differences did not reach statistical significance (P > .10). CONCLUSION: TIPS prior to orthotopic liver transplantation does not shorten operative time and does

not substantially alter intraoperative transfusion requirements. 60 • 2:45 PM Effect of Contrast Material Injections on Portal Venous Hemodynamics M.D. Darcy, MD, St Louis, MO • M.E. Hicks, MD PURPOSE: Pressure measurements and portography are used to assess completion of TIPS. Our goal was to determine if contrast material injections in the portal vein (PV) alter the hemodynamics enough to effect determination of portosystemic gradients (PSGs). MATERIALS AND METHODS: After administration of general anesthesia, two percutaneous transhepatic portal catheters were placed in three dogs. One catheter was a multiple-side-hole catheter placed in the splenic vein for contrast material injection. The other was an end-hole balloon-occlusion catheter in the main PV to provide variable obstruction to PV flow and to measure PV pressures. A transvenous right atrial catheter measured central pressures. Cine angiography was used. Magnitude and duration of changes in PV pressures and PSG were recorded while rates and volumes of splenic contrast material injections were varied. Trials were done with no, partial, or complete PVocclusion. RESULTS: With a patent PV, contrast material injections did not significantly increase PV pressure. With injection rates as low as 6 mL/sec, PV pressure and PSG increased more than 5 mm Hg with partial PV occlusion and more than 10 mm Hg with complete PV occlusion. Pressure spikes lasted 7-8 seconds. With partial PV occlusion, collateral veins opacified at high (but not low) injection rates. CONCLUSION: Elevation of PV pressure and PSG is more pronounced and occurs at lower injection rates as the degree of PV occlusion increases. Pressure spikes are short-lived, but artifactual opacification of collaterals can occur during this time.

1994 SCVIR Meeting Abstracts • 23 Volume 5 Number 1

Monday Mternoon • 6D • Scientific Session 11 Caval Filters MODERATOR:

Marshall E. Hicks, MD, 8t Louis, MO

61 • 1:30 PM Inferior Vena Cavography with and without Selective Renal Venography in Diagnosing Renal Vein Variant Anatomy and Its Relationship to Caval Filter Placement E.S. Malden, MD, St Louis, MO • M.E. Hicks, MD D. Picus, MD • M.D. Darcy, MD • T.M. Vesely, MD D.M. Hovsepian, MD PURPOSE: This study was performed to determine if selective renal venography is more sensitive than cavography in depicting variant renal venous anatomy that changes the optimal location of caval filter placement, and to define infrarenal vena caval dimensions. MATERIALS AND METHODS: One hundred ten consecutive patients referred for IVC filter placement prospectively underwent bilateral iliac venography, cavography, and selective renal venography. Normal anatomy was defined as single renal veins bilaterally and a single IVC. lnfrarenal IVC length and width ~anter?~osteriorand lateral) were determined by usmg a sIzmg catheter. RESULTS: Variant renal venous anatomy was found in 10% of patients at cavography and 33% of patients by means of selective renal vein injection. Detected anomalies included circumaortic renal vein (n = 12) multiple renal veins (n = 24), and duplicated IVC ' (n = 1). Selective renal venography demonstrated vari~nt anatomy not suspected at standard cavograph~ m 26 cases (24%). This variant anatomy changed optImal vena caval filter placement in 18 cases (16%). Average infrarenallength was 7.6 cm (range, 0-12.7 cm). On the anteroposterior view, average infrarenal width wa~ 1.9 cm at the second lumbar vertebral body, 1.9 cm mIdway between the renal veins and caval bifurcation, and 2.0 cm at the caval bifurcation. On the lateral projection, the average widest anteroposterior dimension was 1.8 cm. CONCLUSION: Selective renal venography demonstrates significant variant renal venous anatomy not seen at standard cavography that changes optimal vena caval filter placement in 16% of cases.

62 • 1:45 PM Data Analysis of Vena Cava Filter Placements from the SCVlR Contrast Registry D.R. McFarland, MD*, Little Rock, AR E.J. Ferris, MD* • M.B. Price, BSN* SCVIR Contrast Agent Registry Investigators PURPOSE: Preliminary SCVIR Contrast Registry data were analyzed for patients with IVC filters. MATERIALS AND METHODS: Preliminary data from the multicenter Contrast Registry database are available on 1,847 IVC filter placements. The data were entered into the database at the time of the study. Filter placements were extracted and demographic and risk factors examined. RESULTS: Mean age of patients with filters was 62 years. Seventy-seven percent (n = 1,422) of filters were placed in whites, 16% (n = 296) in Afro-Americans, 5% (n = 92) in Hispanics, and 2% (n = 37) in others. Fifty-one percent (n = 942) of patients were current or past smokers, and 40% (n = 739) had a history of malignancy. Chronic obstructive pulmonary disease was present in 10% (n = 185) of patients, and renal failure was present in 13% (n = 240). Cardiovascular risk factors included thromboembolic events in 91% (n = 1,681), hypertension in 35% (n = 646), prior myocardial infarction in 13% (n = 240), and cerebral vascular accident in 13% (n = 240). Two percent (n = 37) of patients with filters had a history of previous severe allergic reaction. Adverse events were reported in 7.3% (n = 135) of patients and were considered serious in 0.6% (n = 11). CONCLUSION: The SCVIR Contrast Registry serves as a starting point in understanding the current practice of IVC filter placement. 63

2:00 PM

Indications for Placement of IVC Filters K. Kandallu, MD, Chicago, IL • J. Leef, MD J. Rosenblum, MD PURPOSE: We evaluated the indications for vena cava filter placement in a high-volume academic department and assessed how well they conformed to conventionally accepted indications. MATERIALS AND METHODS: A retrospective review of charts and discharge summaries was performed for the 2-year period extending from January 1,1991, through December 31,1992. A total of292 patients underwent placement of vena cava filters. The admission diagnosis and recorded indications for vena cava filter placement were noted. RESULTS: Indications for vena cava filter placement wer~ as follows: recent surgery or prophylaxis for upcommg surgery, 25% of cases; active hemorrhage or hemorrhage immediately following the start of anticoagulation therapy, 20%; concern for either exacerba-

24 • Journal of Vascular and Interventional Radiology January-February 1994

tion of clinically insignificant bleeding or recurrence of previously significant bleeding, 15%; anticoagulation failure, 8%; thrombocytopenia with suspicion of antiheparin AB, 5%; and other causes, 12%. No indication was documented in 15% of cases. CONCLUSION: We found up to 30%-40% of vena cava filters placed at our institution were for other than conventionally accepted indications. We believe this may indicate a need for greater education of our clinical colleagues on appropriate indications. 64 • 2:15 PM Utilization Patterns of IVC Filters: Surgical versus Percutaneous Placement

D.J. Kase, MD, Manhasset, NY • K.S. Crystal, MD M.A. Shapiro, MD • J.B. Naidich, MD PURPOSE: We compared, in terms of numbers of procedures and their indications, changes associated with the shift from surgical to interventional radiologic placement of IVC filters. MATERIALS AND METHODS: In 1989, percutaneous filter placement by radiologists was introduced at our hospital. A retrospective analysis for the years 1986 through 1988 and 1990 through 1992 included the number of filters placed, method of placement, indications for placement, and patient survival. RESULTS: From 1986 through 1988, 34 filters were placed, all by surgeons in the operating suite. From 1990 through 1992,201 filters were placed, exclusively by radiologists in the special procedures suite. The indications were always venous thrombosis with or without pulmonary embolism in patients with contraindications to anticoagulation therapy, or recurrent disease despite anticoagulation. The 6-month survival for the 34 patients treated prior to 1989 was 82% compared with 44% for patients after 1989. CONCLUSION: At our institution, filters are now preferentially placed by radiologists to the exclusion of surgeons. The shift from surgical to radiologic placement was associated with a 591% increase in utilization. Surprisingly, the indications for filter placement remained unchanged. Because more filters were placed in terminally ill patients, survival was reduced. The ease of scheduling and relative ease of insertion accounted for the increased utilization. 65 • 2:30 PM Reassessment of IVC Filter Use in Patients with Cancer

M.P. Rosen, MD, Boston, MA • D.H. Porter, MD D.Kim,MD PURPOSE: Since 1985, the number ofIVC filters placed at our institution has increased from 44 to 82 per year. During this period the incidence of patients dying after filter placement and before or immediately

after hospital discharge has almost doubled from 7.8% to 15.2%. This observation raises questions concerning the appropriateness of filter placement in some patients. The purpose of this study was to define risk factors associated with death soon after IVC filter placement and to develop revised guidelines for filter placement. MATERIALS AND METHODS: During a 4-year period, 141 IVC filters were placed in 137 patients. Patients were divided into two clinical risk groups: presence or absence of malignancy and presence or absence of suprainguinal venous thrombus. Survival was followed for up to 3 weeks after hospital discharge. RESULTS: Death occurred in 16 (26.2%) of 61 patients with malignancy (P = .0086, compared with patients without malignancy); seven (35.0%) of 20 patients with suprainguinal venous thrombus (P = .0422, compared with patients without suprainguinal venous thrombus), and six (46.2%) of 13 patients with malignancy and suprainguinal venous thrombus (P = .0091, compared with patients without malignancy or suprainguinal venous thrombus). CONCLUSION: These data indicate that for some patients with malignancy and/ or suprainguinal venous thrombus, insertion of an IVC filter provides little or no survival benefit. A reassessment of IVC filter use in these patients is warranted. Revised criteria for filter placement will be discussed. 66 • 2:45 PM Outcome of Patients with Advanced Neoplastic Disease Receiving IVC Filters S. V. Lasser, MD, Washington, DC • K.H. Barth, MD PURPOSE: Considering the limited expected survival of patients with advanced malignancies, the appropriateness ofIVC filter placement may be questioned. Therefore, we examined the outcome of filter placements in these patients. MATERIALS AND METHODS: Over a 34-month period, 35 IVC filters were inserted in 34 adult patients with advanced malignant neoplasms. Follow-up was continued for 20 months (median, 5.2 months). Filter effectiveness, complications, recurrent pulmonary emboli, patient survival, and hospital discharge status were recorded. RESULTS: Twenty-eight patients (82%) were discharged home (n = 21) or to nursing facilities (n = 7) between 1 and 193 days (mean, 24 days) after filter insertion. Six patients (18%) died during hospitalization 1-95 days after filter insertion. At the end of the 20-month follow-up period, 12% of patients were still alive. The overall mean survival was 6.6 months; for patients with stage III or IV tumors, mean survival was 8.7 and 5.5 months, respectively. Even among patients with stage IV lesions, 60% survived longer

1994 SCVIR Meeting Abstracts • 25 Volume 5 Number 1

than 3 months. There were no complications related to filter insertion, and no clinical evidence of recurrent pulmonary emboli. In 14% of patients, filters enabled invasive therapeutic and palliative procedures to be performed. CONCLUSION: The presence of relatively advanced neoplastic disease should not be a deterrent to insertion oflVC filters, as the great majority of patients still survived well beyond initial hospitalization.

Monday Mternoon • 6E • Scientific Session 12 Angioplasty MODERATOR:

Kevin L. Sullivan, MD, Philadelphia,

PA 67 • 1:30 PM Peripheral Angioplasty: Extended Follow-up to 10 years A Cotsen, MD, Chicago, IL • R.L. Vogelzang, MD PURPOSE: We extend patency data for peripheral angioplasty to 10 years. MATERIALS AND METHODS: We reviewed results of 126 angioplasties performed on 118 patients between 1979 and 1984. Stenoses or occlusions of 67 aortoiliac and 59 femoropopliteal arteries were dilated. Objective follow-up was obtained at yearly intervals and was continued until occlusion, death, or loss to follow-up. Data were analyzed with Kaplan-Meier life-table survival curves. RESULTS: Five-year cumulative patency was 62% for aortoiliac and 49% for femoropopliteallesions. Tenyear patency was 42% for aortoiliac and 35% for femoropopliteallesions. CONCLUSION: Our 5-year patency rates are comparable with previously published data. Moreover, our la-year patency rates provide important new data that show that balloon angioplasty is an effective, durable procedure which compares favorably with surgical bypass. 68 • 1:45 PM PTA of Hepatic Artery Stenosis after Liver Transplantation: Experience in 20 Allografts

P.D. Orons, DO, Pittsburgh, PA • G.T. Trecha, MD AB. Zajko, MD • R. Sheng, MD • K.M. Bron, MD PURPOSE: We evaluated the safety and efficacy of PTA on allograft survival in the treatment of hepatic artery stenosis after liver transplantation.

MATERIALS AND METHODS: Hepatic artery PTA was attempted in 19 patients (20 allografts). A retrospective analysis was performed to determine if PTA was helpful in prolonging allograft survival. Liver enzyme levels obtained before and after PTA were compared to determine if changes in liver function occurred in response to successful PTA. RESULTS: Technical success was achieved in 16 allografts (80%). Two significant procedure-related complications occurred, including an extensive dissection, which required no treatment, and an arterial leak requiring surgery. Eight allografts (40%) have failed, necessitating retransplantation from 2 weeks to 27 months after presentation. These included the four in which PTA was unsuccessful (P < .05). The indication for retransplantation in these four cases included ischemia in three and rejection in one. Of the four patients with successful PTA who underwent retransplantation, indications included ischemia in one, rejection in two, and cryptogenic cirrhosis in one. Review of liver enzyme levels indicated that marked allograft dysfunction at presentation was associated with an increased incidence of retransplantation as well (P < .05). There was no pattern identified regarding response ofliver enzyme levels to successful PTA. CONCLUSION: PTA of hepatic artery stenosis after liver transplantation is a relatively safe therapeutic modality that may decrease the incidence of allograft loss due to ischemia. However, marked allograft dysfunction at presentation is a poor prognostic sign, indicating the importance of timely intervention. 69 • 2:00 PM Balloon Angioplasty and Thrombolysis of Hepatic Artery Lesions in the Liver Transplant Patient

AS. Gomes, MD, Los Angeles, CA S.C. Goodwin, MD • R.W. Busuttil, MD PURPOSE: Reduction in hepatic arterial blood flow to the liver transplant can severely jeopardize allograft function. Hepatic artery stenosis can result in ischemic biliary tract damage. Severe stenosis or occlusion can result in infarction or loss of the allograft. This report describes our experience with hepatic artery percutaneous balloon angioplasty and thrombolysis and use of recombinant human tissue-type plasminogen activator (rt-PA) or urokinase in the liver transplant patient. MATERIALS AND METHODS: Balloon angioplasty was performed with standard techniques. Intraarterial thrombolysis was performed by using low-dose urokinase with infusion time up to 18 hours; rt-PA was infused at 0.1 mg/kg/h with infusion time of 3-8 hours. RESULTS: Thus far hepatic artery PTA has been attempted in six patients and has been successful in

26 • Journal of Vascular and Interventional Radiology January-February 1994

three with resultant improvement in allograft function. Thrombolysis of the occluded hepatic artery has been attempted in five patients and has been successful with rt-PA in three. CONCLUSION: Percutaneous balloon angioplasty is effective for restoring hepatic blood supply and can result in improved allograft function. Low-profile angioplasty catheters are recommended. Thrombolysis of the hepatic artery can be achieved with use ofTPA; however, this does not always preserve allograft function. This may relate to the duration of allograft ischemia. 70 • 2:15 PM Randomized Double-Blind Trial of Midazolam and Tramadol versus Placebo and Tramadol for Premedication in PTA H. Alfke, Marburg, Germany • H.J. Wagner, MD PURPOSE: Safety and efficacy of two drug regimens for premedication with PTA were studied prospectively. MATERIALS AND METHODS: In a randomized and double-blind fashion, a comparison of midazolam/tramadol versus placebo/tramadol prior to PTA was carried out on 40 patients (12 women, 28 men; mean age, 66.1 years ± 12). The anxiolytic and analgesic effects were quantified by means of a visual analogue score. Hemodynamic responses were monitored, and preand peri-interventional blood gas was measured. The complications of the two schemes were compared. RESULTS: Nineteen patients received midazolaml tramadol, and 21 received placebo/tramadol. Only in the midazolam group was anxiety significantly reduced (from 25.8 ± 25 to 4.3 ± 6). Significant increase in the pain score during PTA was observed only in the placebo group (from 4.3 ± 12.6 to 27.4 ± 20.9). There were no significant changes in hemodynamic data. Respiratory depression was not observed, and the complication rate was comparable. CONCLUSION: The midazolam/tramadol combination seems to be as safe as and more efficacious than tramadol alone as a premedication scheme for PTA.

71 • 2:30 PM Atherosclerotic Lesions Spiraling through the Femoral Artery P.J. Wensing, MD, Utrecht, The Netherlands L. Meiss • W.P. Mali • B. Hillen PURPOSE: Atherosclerosis is common in the adductor hiatus region. The aim of this study was to evaluate atherosclerosis in relation to local morphology in the femoropopliteal region. Two anatomic features are thought to play an important role in the origin of the lesions: (1) the curvature of the vessel, which may lead to unfavorable local hemodynamic factors which

change during leg movements; and (2) the abrupt changes in stiffness of surrounding tissues of the vessel, for instance, the sharp edge of the aponeurosis of the great adductor muscle. MATERIALS AND METHODS: The distal part of 25 femoral arteries was investigated after death. Cross sections were obtained every 0.1 em over a length of 10 em. For every cross section, the lesion thickness was measured at 12 points along the circumference of the vessel. RESULTS: This study revealed no apparent relation between surrounding structures of the femoral artery or the aponeurosis of the great adductor muscle and the location of atherosclerotic lesions. Three-dimensional reconstructions of vessel showed that the atherosclerotic lesions were spiraling through the artery in 20 of 25 cases. CONCLUSION: The spiraling atherosclerotic lesions are coherent with expected flow patterns in this part of the femoral artery. 72 • 2:45 PM Comparison of Cross-Sectional Lumen Area and Stent-induced Neointima Measured by Means of Intraluminal US and Histology T. Takeda, MD, Washington, DC • J. Froelich, MD K.H. Barth, MD • S. V. Lasser, MD D. Lindisch, RT PURPOSE: We determined the ability of 30-MHz intraluminal US to measure histologically verified neointimal buildup and lumen area after stent placement in normal canine iliac and femoral arteries. MATERIALS AND METHODS: The center crosssectional lumen area and the area of neointima formed within 4 months after placement of Palmaz stents and Wallstents in common iliac and common femoral arteries were measured in four large mixedbreed dogs in vivo with a 30-MHz intraluminal US probe (CVIS, Sunnyvale, Calif) and after death by means of corresponding cross-sectional histologic study following in situ vascular fixation under physiologic pressures. Thirteen identical digitized contour tracings and area computations were performed for both modalities. RESULTS: The stents were identified consistently by bright echoes, and any parallel reflective surface detected inside the stent echoes was considered neointima and was found inside all stents in various appearances. The correlation coefficient (r) between intraluminal US and histologic measurements for neointima was r = < .5; for the lumen area it was r = .8. CONCLUSION: High-frequency intraluminal US does not have the resolving power to determine submillimeter neointimal thickness after stent placement

1994 SCVIR Meeting Abstracts • 27 Volume 5 Number 1

in normal iliac and femoral arteries; it does, however, allow measurement of the lumen area with reasonable accuracy.

Monday Afternoon • 6C • Scientific Session 13 MR Angiography MODERATOR:

Edgar K. Yucel, MD, Boston, MA

73 • 3:45 PM MR Venographic Evaluation of Saphenous Veins as Potential Graft Material R.A. Baum, MD, Philadelphia, PA G.A. Holland, MD • J.P. Carpenter, MD R.S. Goldberg • C. Cope, MD PURPOSE: Saphenous veins are commonly used as vascular conduits in the peripheral and cardiovascular systems. Preoperative patency is usually assessed by using Doppler US or contrast venography. The purpose of this study was to determine if MR venography could reliably image the saphenous vein and accurately depict veins suitable for use as bypass material. MATERIALS AND METHODS: Fifteen patients (30 limbs) undergoing coronary artery or peripheral bypass were studied preoperatively with MR venography. MR venography was performed at 1.5 T with two-dimensional time-of-flight technique from the inguinal ligament to the ankle bilaterally. The size and quality (flow and wall characteristics) of each saphenous vein were recorded at the saphenofemoral junction, midthigh, knee, midcalf, and malleolus. At the time of vein harvest, the quality and size of each vein were measured in situ and when distended with saline. RESULTS: Seventeen of 18 veins were considered suitable for use as vascular conduits at MR venography. This was confirmed in every case. One vein found to be recanalized was described at MR venography as having an abnormally thick wall. One patient whose duplex vein mapping did not reveal suitable veins was found to have an adequate saphenous vein at MR venography. This was confirmed at surgery. MR venography measurements of vein size were intermediate between that of the in situ and distended vein and correlated most closely with the size of the distended veIn. CONCLUSION: MR venography is a reliable and accurate method of evaluating the saphenous veins prior to cardiovascular or peripheral vascular bypass surgery.


• 4:00 PM

Propective Evaluation of MR Angiography for the Detection of Renal Artery Stenosis G.A. Holland, MD, Philadelphia, PA A. Chiaramonte, MD • S. Hertz • J. Carpenter, MD C. Cope, MD • R.A. Baum, MD PURPOSE: The purpose of this prospective study was to evaluate MR angiography in the detection of renal artery stenosis (RAS). MATERIALS AND METHODS: Thirty-six patients were studied in a randomized order with contrast angiography and MR angiography to rule out RAS in 72 renal arteries. MR examinations were performed with multicoil imaging at 1.5 T by using two-dimensional time-of-flight (2D TOF) technique in both the axial and sagittal planes with inferior saturation to eliminate venous signal from the infrahepatic IVC (TRITE, 45/7 msec; flip angle, 45°; section thickness, 1.5 mm). Three-dimensional (3D) phase-contrast imaging was also performed with body coil imaging both before and following intravenous administration of gadolinium DTPA at a dose of 0.1 mglkg (TRITE, 24/7 msec; flip angle, 15°; velocity encoding, 30 cm/sec; section thickness, 2 mm). Additional imaging was performed with standard spin-echo technique in the axial plane. MR and contrast x-ray angiograms were read by three radiologists in a blinded fashion and the readings compared. Conventional angiography was considered the standard. RESULTS: RAS was documented at contrast angiography in 14 of the 36 patients studied (38%). MR angiography demonstrated 11 of the stenoses for a sensitivity of 79%. There were no false-positive diagnoses, for a specificity of 100%. Other abnormalities depicted at MR angiography included: renal cysts (n = 8) and renal masses (n = 2). Precontrast 3D phase-contrast and 2D TOF images depicted all stenoses; however, this technique overestimated stenoses by 20%-30%. Postcontrast 3D phase-contrast imaging underestimated RAS by 5%-10%. The three false-negative studies were in the lateral third of the vessels where images were degraded by respiratory artifact. CONCLUSION: MR angiography provides a noninvasive technique to enable diagnosis of RAS especially in the medial two-thirds of the vessel where the majority of atherosclerotic stenoses occur.

28 • Journal of Vascular and Interventional Radiology January-February 1994


4:15 PM

Spiral CT Angiography: Evaluation of Renal Arteries M. T. Farres, Vienna, Austria • W. Dock F. Winkelbauer • R. Wildling • S. Thurnher J. Lammer PURPOSE: The accuracy of spiral CT angiography (a noninvasive modality) was assessed in the evaluation of renal arteries. MATERIALS AND METHODS: With a Somatom Plus S CT scanner (Siemens), spiral CT angiography was performed on 46 patients: 15 patients with renovascular hypertension (before and after PTA), 11 patients after renal artery stent placement, and 20 patients with abdominal aortic aneurysms. The technique included intravenous injection of 120 mL of contrast medium at 3-4 mL/sec; 18-25-second scan delay; 2-3-mm section thickness; 2-6-mm/sec table movement; 210 rnA; postprocessing on maximum intensity projection and surface shaded display. RESULTS: Vessel morphology and stenosis or occlusion of the renal arteries were accurately assessed with spiral CT angiography. The different projections on spiral CT angiography allowed for better evaluation of type of stenosis, location, residual stenosis, stent placement, and calcifications than with conventional angiography. CONCLUSION: Spiral CT angiography is an accurate technique for the evaluation of renal artery morphology. This method can be used as a noninvasive screening technique for renovascular hypertension and as a means of following up PTA results and stent location. 76

4:30 PM

MR Angiography of Renal Transplants G.A. Holland, MD, Philadelphia, PA R.A. Baum, MD • J. Carpenter, MD K. Braemen, MD • M.D. Schnall, MD C. Cope, MD, et al PURPOSE: The purpose of this study was to compare three-dimensional (3D) phase-contrast MR angiography with two-dimensional time-of-flight (2D TOF) MR angiography. MATERIALS AND METHODS: Sixteen patients with renal transplants and clinically suspected renal artery stenosis were examined with MR angiography. Ten patients underwent conventional arteriography. MR angiography was performed at 1.5 T with 2D TOF and 3D phase-contrast techniques acquired with 2-mm-thick axial sections. 2D TOF imaging was performed with the following parameters: TRITE = 33/7 msec, 45° flip angle, one acquisition, and 128 x 256 matrix. 3D phase-contrast imaging was performed with the following parameters: three-direction phase encoding, 30 mml sec velocity encoding, 15° flip angle,

and 128 x 256 matrix. Three radiologists read the MR and conventional angiograms in a blinded fashion. RESULTS: The 3D phase-contrast technique provided the most reliable images of the venous and arterial anastomoses. The 2D TOF images had irregularities noted in the contours of the vessel wall that were not apparent on the 3D phase-contrast images nor on the conventional arteriograms. In addition, the 2D TOF images suffered from saturation of in-plane flow, which compromised the visualization of the anastomoses. MR angiography depicted two external iliac stenoses proximal to the transplanted renal artery (60% and 70%) and two arteriovenous fistula. These findings were confirmed with conventional angiography, yielding sensitivities and specifities of 100% for MR angiography. CONCLUSION: These preliminary results suggest that MR angiography may provide a reliable noninvasive imaging test to screen for vascular disorders affecting renal transplants. 77


Utility of MR Angiography in the Evaluation of Hepatic Artery Anatomy M.R. Middlebrook, MD, Philadelphia, PA A.J. Morillo, MD • G.A. Holland, MD M.D. Schnall, MD, • M.J. Pentecost, MD H. Y. Kressel, MD PURPOSE: We evaluated the ability ofMR angiography to determine hepatic artery anatomy. MATERIALS AND METHODS: Over 3 years, 25 patients underwent diagnostic abdominal MR imaging studies that included three-dimensional phase-sensitive gradient-echo sequences (3DPC) in addition to conventional spin-echo sequences (1.5 T, GE Signa unit). Contrast material-enhanced arteriograms were also obtained. Two skilled observers, blinded to the angiographic results, evaluated the studies. RESULTS: Classic hepatic artery anatomy was found in 19 of 25 patients (76%), but five of these were associated with other variants. These included an occlusion and a separate origin of the splenic artery, narrowing of the celiac origin with reversal of flow in the GDA (n = 2), and origin of the RHA off the GDA. There were replaced RHA (n = 3), replaced LHA (n = 1), replaced LHA and RHA (n = 1), and accessory LHA (n = 1). Overall, 11 patients (44%) h~d variations from the classical anatomy. MR angIograms were considered limited in nine cases, four because of excessive pulsatility artifacts, three because of inadequate vessel signal, and two due to suboptim~l coverage of the vessels within the selected field of VIew. The MR observers each diagnosed 15 of 25 (60%) and 13 of 25 (52%) of the cases correctly. One observer incorrectly identified two classical anatomies as anomalous.

1994 SCVIR Meeting Abstracts • 29 Volume 5 Number 1

More importantly only one of 11 (9%) of the variants were identified by each observer. CONCLUSION: Abdominal 3DPC MR angiography does not yet allow adequate evaluation of hepatic vascularity. 78 • 5:00 PM Magnitude and Range of Image Magnification in Renal and Superior Mesenteric Arteries: Implications for Selection of Balloon Diameter in Renal and Mesenteric PTA S.C. Rose, MD, Seattle, WA • E.J. Verschuyl, MD D. Lopez, RT(R) • W.H. Bush, Jr, MD PURPOSE: With respect to PTA, the likelihood of both restenosis and vascular injury is related to correctly matching balloon diameter to the target artery. Our goal was to determine the degree and range of anticipated magnification that may be encountered in renal and mesenteric balloon angioplasty. MATERIALS AND METHODS: Measurements were made in 50 consecutive adult patients who underwent abdominal CT and included the distance from the origin and distal portions of both renal arteries to the anterior and posterior body surfaces, as well as the origin of the superior mesenteric artery to the lateral body surfaces. Assuming a 40-inch (101.6-cm) filmfocal distance, the degree of magnification was calculated for posteroanterior (PA), anteroposterior (AP), and lateral film techniques. RESULTS: Calculated mean magnification, (range) and film technique for each artery follow. Proximal left renal artery: PA was 15% (range, 100/0-26%), AP 20% (range 16%-26%); distal left renal artery: PA 16% (range, 12%-23%), AP 18% (range, 14%-26%); proximal right renal artery: PA 14% (range, 10%23%), AP 20% (range, 17%-27%); distal right renal artery: PA 16% (range, 11%-27%), AP 18% (range, 14%-24%); superior mesentery artery: left lateral 19% (range, 14%-34%), right lateral 22% (range, 15%-29%). CONCLUSION: The arterial type, portion, and side, and acquisition technique modestly affected the degree of magnification. Variations in body habitus and arterial course substantially affect the range of image magnification. Recommendations will be discussed.

Monday Afternoon • 6F • Scientific Session 14 Biliary Stents MODERATOR:

Anthony C. Venbrux, MD, Baltimore,

MD 79 • 3:45 PM Evaluating Suspected Common Bile Duct Malignancies: A New Biopsy Algorithm P.S. Cook, MD • A.B. Colowick, MD, Denver, CO R.B. Hesky, MD PURPOSE: Suspected malignancies of the biliary tract are difficult to diagnose. We present a multifactorial retrospective review to determine the optimal biopsy algorithm. We also report our experience using a novel technique, an endoluminal percutaneous cutting catheter. Our data suggest a new approach to sampling the biliary tree that results in increased sensitivity and decreased cost. MATERIALS AND METHODS: From 1987 through June 1993, 18 patients with common bile duct malignancies underwent 15 brush biopsies (BBs), 12 fineneedle biopsies (FNBs), and eight percutaneous cutting catheter biopsies (PCCBs) (Simpson atherectomy device). Sensitivity, days to diagnosis, and costs were analyzed. The Fisher exact test was used to compare the three procedures. RESULTS: The sensitivities of the three procedures were as follows: PCCB 88%, FNB 57%, and BB 20%. Both PCCB and FNB were statistically better compared with BB (PCCB, P = .006 and FNB, P = .039). Average days to diagnosis for the PCCB group was 1. 7 and in the non-PCCB group 6.4. Average cost for diagnosis in the PCCB group was approximately 70% of costs for the non-PCCB group. There were no complications. CONCLUSIONS: We advocate that if a mass can be radiologically identified, FNB should be attempted before PCCB; if a discrete mass cannot be identified, PCCB is the procedure of choice. BBs are not costeffective as the initial procedure. 80 • 4:00 PM Factors Affecting Patency and Patient Survival after Biliary Stent Implantation T.K. Roeren, MD, Heidelberg, Germany G.M. Richter, MD, • W. Tonn, BS • M. Brado, MD G. W. Kauffmann, MD PURPOSE: We extract and define demographic, diagnostic, and technical factors from a patient population

30 • Journal of Vascular and Interventional Radiology January-February 1994

with malignant biliary obstruction that may influence patency of metallic biliary stents and patient survival. MATERIALS AND METHODS: In a prospective study from 1989 to 1993, 79 patients received metallic biliary stents for malignant biliary obstruction. Patients were followed up on a regular basis. Stent patency was controlled by means of clinical, laboratory, and sonographic parameters. Survival after stent placement and stent patency were correlated with sex, age, type and stage of underlying tumor, and concomitant therapies. Additionally, stent type (Wallstent, Palmaz, coated Palmaz, Strecker), length, and diameter were correlated. Significance level was P < .05. RESULTS: With a primary patency of 85%, the various factors investigated affected patency and survival to an almost equal extent. Overall survival was not significantly affected by tumor histology and tumor staging, although patients with pancreatic cancer had a tendency to survive longer than patients with cholangiocarcinoma (160 vs 98 days). Patients who underwent chemotherapy lived significantly longer (350 vs 139 days); patients who underwent radiation therapy did not profit (124 vs 175 days). Stent length and diameter were without influence. Depending on stent type, patients had a mean survival ranging from 140 (Wallstent) to 207 days (coated Palmaz), but the difference is not significant. Patients with distal tumors had an almost significantly longer survival than those with proximal tumors. CONCLUSION: Our prospective study group is rather large for the type of disease investigated, but out of various factors influencing the clinical course, to date the only significant factor affecting survival and stent patency is concomitant chemotherapy. Our analysis reveals strong tendencies in favor of distal tumors (mostly pancreatic cancer). The preference for Wallstents when the hepatic bifurcation is involved explains the shorter survival times for this patient group. 81 • 4:15 PM Use of Original, Modified, and Spiral Z Stents in the Biliary System H. Uchida, MD, Nara, Japan • T. Yoshioka, MD H. Sakaguchi, MD • M. Maeda, MD

K. Kichikawa, MD PURPOSE: We evaluated the clinical efficacy of Z stents for the treatment of malignant biliary obstruction, especially bile duct carcinoma (BDC). MATERIALS AND METHODS: Since 1987, 164 patients with malignant biliary obstruction were treated by using original, modified, and spiral Z stents. Eighty-nine patients had biliary obstruction due to BDC; in 61 of these radiation therapy (RT), including external and internal RT, was performed before stent

placement. Tube-free, patency, and survival rates were analyzed. RESULTS: The external drainage tube could be removed in 140 of 164 patients (84.5%). The survival rate of patients with BDC was 32% at 1 year, 13% at 2 years, and 2% at 3 years after tube removal following stent placement. The cumulative survival rate was higher in patients with BDC receiving RT compared with those in the non-RT group. Although the spiral Z stent has several advantages, such as flexibility and expansile force, no significant difference was seen in therapeutic effect among the three types of Z stents. CONCLUSION: Z stents are very useful and play an important role in achieving a better quality oflife in patients with malignant biliary obstruction and the combination of Z stent and RT seems to be effective for patients with BDC. 82 • 4:30 PM Long-term Dilation of Chronic Biliary and Ureteral Stenosis with Use of a Retrievable Wallstent Incorporated into a Drainage Catheter H. Bjarnason, MD, Minneapolis, MN

G.K. Nazarian, MD • D.J. Stackhouse, MD D. W. Hunter, MD • K. Amplatz, MD PURPOSE: The purpose of this animal study (canines) was to test the feasibility of using Wallstents incorporated into a drainage catheter for long-term dilation of ureteral and biliary stenoses. MATERIALS AND METHODS: Noncoated and silicon-coated Wallstents were incorporated into a lO-F drainage catheter and were placed operatively through the gallbladder into the common bile duct. The stent was left in for 1-4 weeks. The ease of removal and the changes in the bile duct wall were evaluated. In one animal, a stenosis was created surgically in the common bile duct, and the effect of long-term dilation was tested. RESULTS: In the first part of the study, two noncoated stents were used. These became embedded in the endothelium, and the stents were not retrievable at 1 week. Three coated stents were placed, and none became embedded in the endothelium. All were easily retrieved for up to 4 weeks. CONCLUSION: Our results indicate that the technique oflong-term dilation and drainage of ureteral and biliary stenoses is feasible. We believe this method may improve the success rate for the treatment of these difficult stenoses. Possible application in malignant stenosis is of interest. Further animal work is in progress.

1994 SCVIR Meeting Abstracts • 31 Volume 5 Number 1

83 • 4:45 PM Placement of Covered Self-expanding Stents in the Common Bile Duct: A Feasibility Study K.C. Wright, PhD, Houston, TX • K. Yasumori, MD N. Mahmoudi, MD • S. Wallace, MD C. Gianturco, MD PURPOSE: Metallic stents have not worked well in the presence of intraluminal tumor because of tumor growth between the wires. Self-expanding stents were covered with polymeric membranes, which may prevent tumor ingrowth, and the feasibility of using these covered stents in the common bile duct was evaluated in dogs. MATERIALS AND METHODS: Initially, stents (10 x 6 mm) covered with polycaprolactone (PCL), segmented polyether polyurethane (SPP), and silicone were placed and followed up for 3-12 months. Subsequently, four types of modified silicone-covered stents were evaluated. Rough-surface, small-diameter (3 mm), half-covered, and tapered stents were placed and followed up for 3 months. RESULTS: In the initial study, the SPP and silicone membranes remained intact, but the PCL covering broke in all dogs. In all but two dogs, luminal narrowing caused by varying degrees of papillary mucosal hyperplasia was observed at both ends of the stent. In the subsequent study, mucosal hyperplasia was considerably reduced by using half-covered stents and was virtually eliminated by tapered stents. All smalldiameter stents migrated into the bowel. CONCLUSION: These studies suggest that tapered self-expanding stents covered with silicone may be useful in the biliary tract. 84 • 5:00 PM Midterm Results of Expandable Metallic Stents in the Treatment of Biliary Strictures T. Kimoto, MD, Fukui, Japan • N. Hayashi, MD T. Sakai, MD • N. Katoh, MD • S. Murashima, MD Y. Ishii, MD PURPOSE: We summarized the midterm results with Gianturco expandable metallic stents in the treatment of benign and malignant biliary strictures. MATERIALS AND METHODS: Twenty-five patients with obstructive jaundice were treated with Z stents. Patients were aged 49-85 years (mean, 72.9 years). Causes for obstruction included bile duct cancer (n = 13), gallbladder cancer (n = 6), lymph node metastasis at porta hepatis (n = 3), and benign postoperative or inflammatory strictures (n = 3). Either original or Rosch-modified stents were placed through PTCD routes. All the patients were followed-up for 1-44 months. RESULTS: Stent placements were technically successful in every patient. In 24 patients, external drain-

age tubes were withdrawn within a week, while in the remaining one patient, passage through the stents was very poor, and a plastic stent was inserted within the metallic stents. In 22 patients with malignant obstructions, nine patients had recurrent obstructive jaundice before death 0.5-7 months (mean, 4.0 months) after stent placement, and six of them were treated with external biliary drainage. In three patients with benign strictures, one had obstructive jaundice, but additional stent implantation was performed successfully. The other two patients are alive without jaundice 18 and 44 months after stent placements. CONCLUSION: Transhepatic placement of metallic stents is safe and effective. It can be a good palliation for patients with malignant biliary obstruction. Indication for benign disease should be deliberated, but the results are encouraging.

Monday Afternoon • 6D • Scientific Session 15 Vascular Access MODERATOR:

Matthew A. Mauro, MD, Chapel Hill,

NC 85 • 3:45 PM Radiologic Placement of Long-term Central Venous Peripheral Access System Ports M.J. Foley, MD, Tampa, FL PURPOSE: The PAS Port catheter is a peripherally implantable central venous access port that can be safely placed by interventional radiologists. Radiologic and surgical complication rates as well as radiologic and surgical hospital charges are compared. MATERIAL AND METHODS: Results in 150 patients in whom the PAS Port catheter was implanted were retrospectively reviewed. The age range was 26-84 years (mean, 46 years). The diagnoses in these patients included cancer (n = 60), AIDS (n = 52), osteomyelitis (n = 21), abscess (n = 9), and pain management (n = 8). Implantation with fluoroscopic guidance or "cath finder" technology is described. RESULTS: One hundred fifty ports have been indwelling for a total of24,150 patient days (range, 40220 days; mean duration, 161 days). Sepsis occurred in 3%, local infection 5%, catheter occlusion 3.3%, postinsertion phlebitis 7%, and venous thrombosis 2%. These rates are as low or lower than quoted surgical rates. The cost of PAS Port insertion in radiology averaged $1,488 versus $3,720 for surgery.

32 • Journal of Vascular and Interventional Radiology January-February 1994

CONCLUSION: PAS Port central venous catheter placement by interventional radiologists is safe and cost-effective with complication rates at least as good as those associated with surgical placement. Radiologists will likely see increasing requests by referring physicians for radiologically placed catheters and ports in this era of cost containment. 86 • 4:00 PM Radiologic Placement of Long-term Central Venous Access Catheters in a Pediatric Population T.C. McCowan, MD, Omaha, NE T.C. Goertzen, MD • R.P. Lieberman, MD R.F. LeVeen, MD PURPOSE: We evaluated the utility and safety ofradiologic placement oflong-term central venous access catheters in pediatric patients. MATERIALS AND METHODS: Thirty-four attempts at radiologic placement of tunneled long-term central venous access catheters in 19 pediatric patients (age, 60 days to 17 years; mean, 8 years) were reviewed. The site of the access attempts was IVC (n = 23) and subclavian vein (n = 11). Catheter sizes ranged from 7 F to 14.4 F. Insertion indication was therapy for malignancy in 24 of 34 attempts (71%). RESULTS: For IVC catheters, 22 of 23 (96%) insertions were successful. One immediate complication occurred: retroperitoneal hematoma. Catheter dwell time was 8-185 days; mean, 68 days. Catheters were removed because of thrombus (n = 1), infection (n = 7; none < 14 days, three with negative cultures) and completed therapy (n = 4). Eight catheters remain in place. Two patients were lost to follow-up. For subclavian catheters, 11 of 11 (100%) insertions were successful. There were no immediate complications. Catheter dwell time was 8-149 days; mean, 42 days. Catheters were removed because of infection (n = 2, one < 14 days, one with negative cultures) and completed therapy (n = 1). Eight catheters are in place at follow-up. CONCLUSION: Success and complication rates in this radiologic pediatric series are comparable with reports of surgically and radiologically placed longterm central venous access catheters in adults. 87 • 4:15 PM Peripheral Insertion of Central Venous Catheters Guided by Means of Carbon Dioxide Contrast Agent J. Bonn, MD, Philadelphia, PA D.J. Eschelman, MD • K.L. Sullivan, MD G.A. Gardiner, Jr, MD PURPOSE: Carbon dioxide gas was examined as a venous contrast agent to guide upper arm placement

of peripherally inserted central venous catheters (PICCs). MATERIALS AND METHODS: PICCs were placed in 52 consecutive patients for long-term intravenous antibiotic therapy (n = 42), hyperalimentation (n = 5), hydration (n = 2), chemotherapy (n = 2), and blood products (n = 1). Access was obtained in upper arm veins by using fluoroscopic guidance during more peripheral intravenous injection of CO 2 gas. Iodinated contrast material was used only if CO 2 -related symptoms or inadequate visualization prevented completion. Patients were questioned after the procedure about their symptoms during CO 2 injection; angiographers were questioned about the quality of guidance with CO 2 • RESULTS: All PICC insertions were successful without complication, including 27 (52%) with CO 2 guidance alone. Nineteen patients (37%) reported no CO 2 related symptoms; 23 patients (44%) reported moderate or severe burning or pain in the arm or at the peripheral injection site. Three angiographers reported good, very good, or excellent visualization of the target vein with CO 2 in 38 procedures (73%). CONCLUSIONS: CO 2 gas is a safe and effective contrast agent for the fluoroscopic guidance of upper arm venous access for PICC placement. Many patients may still require iodinated contrast material to improve target vein visualization or to reduce symptoms. Further study is planned of the factors causing localized symptoms during venous CO 2 injection.

88 • 4:30 PM Mechanism of Positional Dysfunction of Subclavian Venous Catheters D. W. Hunter, MD. A.E. Krutchen, MD J.A. Parsons, PhD· G. Nazarian, MD, Minneapolis, MN • J.E. Magney, MS PURPOSE: We define the mechanism whereby percutaneously placed subclavian venous catheters malfunction when the patient assumes certain positions. The presumed etiology is crimping of the catheter as it passes through the subclavius muscle or costoclavicular ligament based on a cadaver study of percutaneously placed pacemaker leads. MATERIALS AND METHODS: Nine patients referred to the radiology department with a diagnosis of positional catheter dysfunction were studied with cine during arm and head movements. Once malfunction had been proved, the catheter was replaced with one that entered the subclavian vein in a more lateral position, bypassing the subclavius muscle. RESULTS: All nine patients had clearly documented pinching of the catheter by the musculotendinous tissues and lateral to the catheter's entry into the vein. No patient had compression of the catheter between the clavicle and the first rib. All patients had complete

1994 SCVIR Meeting Abstracts • 33 Volume 5 Number 1

elimination of the catheter malfunction by puncturing into the subclavian vein over the first rib. CONCLUSION: Subclavian catheter positional malfunction is not due to first rib-clavicle compression but to binding of the catheter in the subclavius muscle-costoclavicular ligament complex. Fluoroscopically guided puncture is a very easy, reliable technique for avoiding this problem. 89

4:45 PM

Percutaneous Thoracic Duct Cannulation C. Cope, MD, Philadelphia, PA • K.R. Kensey, MD PURPOSE: The purpose of this ongoing feasibility study is to develop a safe, reliable percutaneous technique for catheterization of the thoracic duct (TD) at the jugular-subclavian venous angle. Because the competent valves of the distal TD make its percutaneous retrograde catheterization from the vein impossible, access to the TD for sampling and draining lymph in experimental and clinical situations is still performed today by means of surgical cannulation or TD-venous anastomosis, with associated potential complications of sepsis, pneumothorax, and persistent lymphocutaneous fistulas. MATERIALS AND METHODS: The TD was opacified at lymphangiography in six large dogs. RESULTS: Attempts to selectively catheterize or puncture the distal TD under fluoroscopy in the neck were unsuccessful in four dogs because the TD broke up into five inaccessible branches before reaching the veins. A procedure was then developed in two animals for catheterizing the TD in an antegrade direction through the cisterna chyli, which permits capture and successful pullback of the venous catheter into the TD. CONCLUSION: This promising procedure is only possible in animals with large cisternae and an uninterrupted TD. 90

5:00 PM

Embolization in Congenital Heart Disease J.F. Reidy, London, England • SA Qureshi M. Tynan • E.J. Baker • I.C. Huggon PURPOSE: The techniques and results of embolization are reviewed in congenital heart disease (CRn). MATERIALS AND METHODS: Two hundred seven patients with CRD underwent embolization. Eightyfive percent of the patients underwent ductus occlusion with Rashkind device; the remaining 30 patients (age range, 7 weeks to 67 years; mean, 11.5 years) underwent 44 embolizations for coronary arteriovenous fistulas (13 arteries in 10 patients), Blalock-Taussig shunts (seven patients), multiple aortopulmonary collaterals in pulmonary atresia (18 arteries in seven patients), ductus arteriosus (four patients), Scimitar

syndrome (one patient), and massive hepatic hemangioma (one patient). Techniques used for the 44 embolizations included nine detachable balloons (n = 9), 73 0.038-inch coils (n = 11), four 0.025-inch coils (n = 2),27 0.018-inch coils (n = 11), 18 interlocking detachable coils (n = 7), and Gelfoam (n = 4). RESULTS: Complete occlusion resulted in all except four lesions, including two aortopulmonary collaterals not satisfactorily catheterized. No significant complications occurred. Five coils and one balloon were misplaced but these events were not associated with clinical problems: four coils were satisfactorily retrieved, and the remaining coil and the balloon (distal pulmonary artery) were asymptomatic and not retrieved. CONCLUSION: Embolization in CRD is safe, and the development of new techniques has made it even more effective. It offers a more precise alternative to major surgery with less morbidity.

Monday Mternoon • 6E • Scientific Session 16 Interventional Radiology Practice MODERATOR:


Albert A. Nemcek, Jr, MD, Chicago, IL

• 3:45 PM

A Macintosh-based Patient Management Program for Cardiovascular/Interventional Radiology H. Bjarnason, MD, Minneapolis, MN D. W. Hunter, MD • R.A. Jarvis D.J. Stackhouse, MD • G.K. Nazarian, MD K. Amplatz, MD PURPOSE: A MacIntosh-based program has been created that allows technologists and physicians to perform a variety of functions related to patient care and quality assurance. MATERIALS AND METHODS: Scheduling is done easily by using pull-down lists and standard data entry. Fields and data lists have been created to facilitate specific searches for quality assurance or research purposes. Special search algorithms have been designed for the most common search needs. A special form and a printing algorithm have been designed to print a procedure note directly from the data selected from pull-down lists or entered in the patient record form on the computer. This has eliminated any "double-entry" problems and resulted in immediate acceptance of the program by the interventional radiology staff as well as the clinical services. RESULTS: With this program, the entire cardiovascular and interventional radiology operation, which

34 • Journal of Vascular and Interventional Radiology January-February 1994

now requires extensive outcome analysis and quality assurance documentation, has been streamlined significantly, with marked improvement in efficiencies of both the technical and medical staff. CONCLUSION: This program has proved to be very user friendly, reliable, and easy to learn. Data retrieval, both for research and quality assurance, is easy. Further development, including follow-up and research functions, is planned. 92

4:00 PM

Computer Program for Documentation of Vascular and Interventional Procedures and Biopsies J.E. Parker, MD*, Tacoma, WA • M.F. Haggerty, MD PURPOSE: We evaluate the use and general acceptability of a new database program with a training program in a radiology department. MATERIALS AND METHODS: At the beginning of a 6-month trial period, residents and staff were given a brief orientation on the operation of the computer program. The program features (1) customized "popup" menus for data entry that reduce entry time and errors; (2) on-screen "browsing" of sorted and filtered data; (3) reports individualized by department section, staff, and assistant; (4) detailed complication reports; and (5) biopsy reports with pathologic findings. RESULTS: Compliance with logging of procedures was approximately 85% the first month, then rose to over 97% after distribution of the first month's reports. Data entry errors declined after the first month, partly in response to customization of the menu lists. Acceptance of the program was enthusiastic by residents and staff. CONCLUSION: The program simplified and centralized documentation of all procedures performed in our department. Availability of accurate printed reports was cited as the main advantage of the program by physicians and other quality assurance personnel. 93

4:15 PM

The Interventional Radiology Clinic: An Improvement in Patient Care and Efficiency H.B. D'Agostino, MD, San Diego, CA D. Johnson, RN • B. Stainken, MD • A. Beale, MD R. o 'Laoide, MD • C. Pittman, MD PURPOSE: The operational details, benefit, and impact of an outpatient nonvascular interventional radiology clinic are reported. MATERIALS AND METHODS: In a 13-month period, 176 patients were seen at the clinic (79 for preprocedure evaluation and 97 for follow-up). Preprocedure evaluation consisted of chart, film, and laboratory test review. A limited physical examination

was performed, coagulation tests were ordered, and anesthesia or admission arrangements were made if considered appropriate. Consent was obtained, and the case was scheduled. Follow-up patients were seen for catheter management or for assessment of treatment results. The effect of the clinic on patient care and interventional service efficiency was measured by the number of procedures canceled because of inappropriate indication, inadequate preparation, number of procedural complications diagnosed in the clinic, and number of unexpected patients who came to the hospital for catheter management problems interfering with daily scheduled activities. RESULTS: Preprocedure consultation allowed recognition of inappropriate procedure indication in two patients. There were no cancellations due to poor patient preparation. Scheduled follow-up permitted the diagnosis of abscess recurrence (four patients), and diagnosis and treatment of severe urosepsis (one patient). Unexpected catheter exchange or manipulation occurred in two patients. CONCLUSION: The outpatient clinic offers benefits for the practice of nonvascular interventional radiology by improving patient care and service efficiency. The outpatient clinic is a step forward toward total patient management by interventionalists. 94


Combined Fluoroscopic and CT Guidance for Interventional Procedures A.J. Cook, MD, Madison, WI • J.C. McDermott, MD A.B. Crummy, MD PURPOSE: Three hundred forty-five patients have undergone interventional procedures performed with dual guidance: CT for lesion localization and fluoroscopy for real-time biopsy or drainage. The feasibility, efficacy, applicability, and safety of dual guidance are discussed. MATERIALS AND METHODS: Procedures were performed with a CT scanner and portable C arm. The fluoroscopic unit was placed between the CT table and gantry, which had modified table-positioning software. A total of 345 procedures were performed on selected patients, and 60 consecutive cases were reviewed retrospectively. Biopsy or drainage was performed under fluoroscopic control after CT localization. RESULTS: With real-time control and tandem needle localization, large-core biopsy samples were confidentlyobtained. Tissue sufficient for diagnosis was obtained from all 44 biopsies. All 16 abscess drainages reviewed were successful. There were no complications of biopsy or drainage procedures, excluding pneumothoraces. CONCLUSION: The combination of CT (superb localization) and fluoroscopy (real-time guidance) in one

1994 SCVIR Meeting Abstracts • 35 Volume 5 Number 1

imaging unit facilitates safe and expeditious interventional procedures. This combined approach increases the range oflesions and fluid collections for which biopsy and drainage can be performed safely.

95 • 4:45 PM Major Complications of Nonvascular Interventional Radiology Techniques: Review of 2,749 Procedures B.F. Branstetter, MD, San Diego, CA H. D'Agostino, MD • R. O'Laoide, MD S. Ogleuie, MD • M. Middleton, MD E. uanSonnenberg, MD PURPOSE: The aim of this presentation is to report the rate of major complications in a series of 2,749 nonvascular interventional radiology (lR) procedures performed over 24 months at the University of California, San Diego (UCSD) IR service. MATERIALS AND METHODS: From January 1991 to December 1992, the UCSD IR service performed 2,749 procedures on 994 patients. Procedures included were drainage (n = 966), biliary (n = 528), urologic (n = 450), biopsy (n = 400), gastrointestinal (n = 285), sclerosis (n = 110), and miscellaneous (n = 10). Procedures were divided into major (n = 1,967) and minor (n = 782). Major complications were defined as those requiring an additional procedure, extended hospitalization, blood transfusion, or intravenous antibiotics. RESULTS: There were 47 complications (1. 7%), including four postprocedure deaths (0.1%). Major procedures were associated with 42 complications (2.1%). Minor procedures were associated with five complications. The complications were pneumothorax (chest and neck procedures, n = 28), septic reaction (catheter manipulations and nephrostomy, n = 4), hemorrhage (chest and urologic procedures, n = 3), perforated viscus (n = 3), respiratory arrest (n = 2), cardiac arrest (n = 1), drug reaction (n = 1), and pneumomediastinum (n = 1). Postprocedure deaths occurred in three critically ill patients, and one morbidly obese patient died from general anesthesia. CONCLUSION: The low rate of complications in this series documents the safety of percutaneous techniques. Patients' poor general condition and inappropriate indications for anesthesia determined mortality in our series. Pneumothorax was the most common major complication encountered.

96 • 5:00PM Thrombolysis of Peripheral Embolic Occlusions: Results of the STAR Registry E.A. Huettl, MD, Philadelphia, PA M. C. Soulen, MD PURPOSE: Thrombolysis was evaluated as primary therapy for lower extremity embolic occlusions. MATERIALS AND METHODS: We reviewed 306 consecutive cases of lower extremity arterial occlusive lesions treated with urokinase and registered in the SCVIR Transluminal Angioplasty and Revascularization (STAR) Registry to evaluate outcome in patients with embolic occlusions. RESULTS: Forty-five patients with embolic occlusions were identified. Comorbidity included histories of myocardial infarction in 40%, atrial fibrillation in 41%, and a cerebrovascular event in 35%. Thirty-two limbs (71%) were viable, 12 (27%) were threatened, and one had irreversible ischemia. Mean symptom duration was 8.6 days; 53% of patients were treated within 48 hours of symptom onset. Average occlusion length was 17 cm. The distribution of emboli was aortoiliac 6%, femoropopliteal 60%, tibial 29%, and graft 5%. A mean of 1.98 million U of urokinase was administered. Mean ankle-brachial indexes before and after the procedure were 0.20 and 0.86, respectively. Major complications occurred in six patients (13%). The initial technical and clinical success rates were 69% and 67%. Life-table analysis yielded a 1- and 2-year unassisted patency of 77%. CONCLUSION: Thrombolysis of embolic occlusions appears to have a lower success rate than the 81% average reported for thrombotic occlusions.

Wednesday Mternoon· 6C • Scientific Session 17 Mechanical Thrombectomy MODERATOR:

Gordon K. McLean, MD, Pittsburgh,


97 • 1:30 PM Evaluation of Endothelial Damage Caused by a Percutaneous Mechanical Thrombolytic Device A. Lajuardi, MD S.O. Trerotola, MD, Indianapolis, IN J.D. Strandberg, DVM, PhD • M.A. Samphilipo, BS C.Magee,BS PURPOSE: Recently, mechanical thrombolysis has received considerable attention as a viable alternative

36 • Journal of Vascular and Interventional Radiology January-February 1994

to urokinase for clot dissolution; however, concern has been raised regarding potential endothelial injury. The purpose of this experiment was to study the effects of a mechanical thrombolytic device (the Trerotola percutaneous thrombolytic device [TPTDJ) on the venous endothelium. MATERIALS AND METHODS: Simulated thrombolysis was performed by using the TPTD in the IVC offorty New Zealand White rabbits. The Fogarty balloon catheter (FBC) was used as the control. Pre- and postprocedure as well as necropsy cavograms were obtained acutely, at 1 week, and at 6 weeks. Rabbits were injected with Evans blue dye prior to death to stain denuded regions. Histologic examination was also performed. RESULTS: There was significantly greater reendothelialization in the distal IVC in the rabbits treated 6 weeks earlier with the TPTD (P :s; .04). Additionally, the inner luminal diameter at necropsy for the I-week rabbits treated with the FBC was significantly narrower in the distal and middle sections of the IVC compared with the TPTD (P :s; .02 for both segments). CONCLUSION: Based on a rabbit model, venous endothelial damage was found to be less with the TPTD than the routinely used FBC. 98

• 1:45 PM

Radiographic and Histopathologic Evaluation of the Vessel Wall after Use of the Amplatz Aspiration Thrombectomy Device M.R. Gomes, MD, Minneapolis, MN G.H. Pozza, MD • Z. Qian, MD R. Ambrozaitis, MD • J. Titus, MD K. Amplatz, MD PURPOSE: The authors evaluated the acute and long-term effects of the Amplatz aspiration thrombectomy (AAT) device on the arterial and venous system. MATERIALS AND METHODS: Forty-two vessels (21 arteries and 21 veins) in 26 mongrel dogs were used. Postprocedure angiography was performed, and the dogs were killed either acutely or after 3 months. Thirty vessels were examined histopathologically. RESULTS: No radiographic damage was present in 35 vessels. Minor extravasation was seen in three vessels, minor stenosis in three, and a filling defect in one vessel. Three-month follow-up showed no abnormalities, except in one case with moderate stenosis. Histopathologic examination in acute cases showed focal intimal disruption in 16 vessels, and normal findings in eight. Examination at 5 and 10 days demonstrated only minor valve damage and focal intimal congestion. At 3 months the vessels were normal. CONCLUSION: Although this device causes focal areas of intimal damage, as proved histopathologically, in the long-term follow-up, no vessel wall abnormali-

ties were identified, due to intimal regeneration. As with any catheterization, the risk of perforation must not be ignored. 99 • 2:00 PM In Vivo Evaluation of the Safety to the Vessel Wall of the Hydrolyser Hydrodynamic Thrombectomy Catheter G. V van Ommen, MD, Maastricht, The Netherlands E. van der Veen • M. Daemen • J. Habets H.J. Wellens PURPOSE: The authors evaluated the safety of the Hydrolyser in vivo. MATERIALS AND METHODS: The Hydrolyser is a recently developed 7-F double-lumen hydrodynamic thrombectomy catheter. With use of a conventional contrast material injector, saline is injected (3 mL/sec, 500-600 psi) into a narrow supply channel that makes a 180 curve at the tip. The high-velocity jet (± 100 mph) is directed over a side hole into a wide discharge channel. Due to the resultant Venturi effect, the thrombus is sucked into the side hole, fragmented, and removed through the discharge channel. In recent animal experiments and 50 clinical cases the efficacy in removing intravascular thrombus material up to 14 days old was proved. To evaluate the effect of this thrombectomy procedure on the vessel wall, the passage of a nonfunctioning (NF) Hydrolyser in the proximal part of a vessel was compared with that of a functioning (F) Hydrolyser in the distal part of the same vessel. Histologic evaluation was performed in three goats (15 normal arteries and 14 normal veins) 10 days after treatment. RESULTS: 0

Arteries Veins


Endothelial Damage 2 1 1 1

Intimal Hyperplasia (Cell Layers) 1-5 0 >5 8 7 0 8 6 1 10 3 1 2 12 0

CONCLUSION: This study shows that the effect on the vessel wall (endothelial reaction) by the thrombectomy action (jet) ofthe Hydrolyser is not different from that of the nonfunctioning system (catheter only).

1994 SCVIR Meeting Abstracts • 37 Volume 5 Number 1

100 • 2:15 PM Percutaneous Mechanical Thrombectomy S.M. Tadavarthy, MD, Minneapolis, MN K. Amplatz, MD* • S. Inampudi, MD P.D. Murray, MD PURPOSE: We present our human experience of mechanical thrombectomy with the Amplatz device in arterial grafts, native arteries, and veins. MATERIALS AND METHODS: The device consists of an 8-F catheter with an axial impeller that rotates at 150,000 rpm, and the thrombus is pulverized. Occasionally it was supplemented with lytic agents. RESULTS: The device has been used in 18 patients: 11 with arterial grafts, five in native arteries, and two in veins. Declotting is achieved in 1-4 minutes, with the underlying stenosis being corrected with PTA or endarterectomy. Complete success was achieved in 12 patients, partial success was achieved in three, and failures occurred in three patients. Overall complications were noticed in 28% of the patients. CONCLUSION: The device cleared thrombi without significant complications and it appears to be faster than fibrinolytic agents. 101 • 2:30 PM Percutaneous Mechanical Thrombectomy of Hemodialysis Shunts with Use of the Hydrolyser System D. Vorwerk, MD, Aachen, Germany R. W. Guenther, MD • M. Sohn, MD K. Schuermann, MD • Y. Hoogeveew, PhD PURPOSE: A new mechanical thrombectomy system utilizing the Venturi effect for thrombus removal (Hydrolyser) was applied for shunt occlusions. MATERIALS AND METHODS: In eight patients with recently thrombosed hemodialysis fistulas, the Hydrolyser catheter was used for thrombectomy. Three patients had Goretex loop shunts, and five patients had native forearm shunts. The catheter was inserted via a 7-F sheath after retrograde venous puncture in native shunts. Imaging was performed via arterial fine-needle access. In Goretex loops, the shunt was cannulated in the antegrade and retrograde direction for thrombectomy of both limbs. Arterial anastomosis was protected by a Dormia mini-basket, and venous embolization was prevented by a tourniquet. RESULTS: Arterialized flow was sufficiently established in six of seven cases. Thrombectomy was almost complete by three to four cycles with the Hydolyser alone but was always combined with PTA for underlying stenosis (n = 4) or residual thrombus (n = 3). Thrombus removal was complete in three of three Goretex loops and three of four native shunts. Residual thrombus did not compromise shunt patency in

one patient but led to early reocclusion in another. No major complications or embolization occurred. CONCLUSION: The Hydrolyser system seems promising for percutaneous thrombectomy of hemodialysis fistulas and may be an alternative to thrombolysis. 102 • 2:45 PM

Central Venous Thrombosis: Treatment with the Amplatz Thrombectomy Device R.D. Edwards, FRCR, Liverpool, England P.C. Rowlands, FRCR • J. Meaney, FRCR PURPOSE: SVC obstruction can be effectively treated with endovascular metal stents, but the presence of extensive central venous thrombosis precludes their use. Thrombolytic therapy may be effective, but a number of patients fail to respond to treatment. To assess the efficacy of mechanical thrombectomy, the Amplatz thrombectomy device (ATD) was used in five patients with central venous thrombosis associated with SVC obstruction. MATERIALS AND METHODS: Five patients with bronchial carcinoma had SVC obstruction, and cavography demonstrated central venous thrombosis. Thrombolytic therapy was employed in four patients without success (mean lysis time, 31 hours). Thrombolysis was contraindicated in one patient. The ATD was introduced via the femoral, jugular, or arm veins and was activated for a mean duration of 5.2 minutes. Following thrombectomy, 50 mL of blood was aspirated from the SVC and endovascular stents were placed in all cases. RESULTS: Clot dissolution was successful in all cases and no clinically significant pulmonary embolism occurred. A complete clinical response was achieved in four cases following stent insertion, and partial resolution of symptoms occurred in one patient. CONCLUSION: Mechanical thrombectomy is an effective alternative therapy in patients with central venous thrombosis when thrombolysis is ineffective or contraindicated.

38 • Journal of Vascular and Interventional Radiology January February 1994

Wednesday Mternoon· 6F • Scientific Session 18 Vascular Diagnosis MODERATOR:

M. Victoria Marx, MD, Ann Arbor, MI

103 • 1:30 PM Incidence of Hematomas in Contemporary Practice of Vascular and Interventional Radiology: Results of the SCVIR Contrast Registry

S. V. Lossef, MD, Washington, DC • K.H. Barth, MD SCVIR Contrast Agent Registry Investigators PURPOSE: The authors attempted to determine the current incidence of all puncture site hematomas from diagnostic and therapeutic arterial and venous procedures. MATERIALS AND METHODS: The incidence of hematomas was determined for 74,496 diagnostic angiograms and vascular interventional procedures included in the CIRREF Contrast Registry. RESULTS: The incidence of minor and major hematomas for diagnostic extremity (2.50%), renal-visceral (2.33%), and cerebral (2.42%) arteriography did not differ significantly (P = .746). For coronary arteriography, the incidence of hematoma (2.97%) was slightly higher (P < .01). Compared with diagnostic arteriography, there was an increased incidence of hematomas (P < .001) for angioplasty of extremity (6.90%), renalvisceral (4.90%), and coronary (4.35%) arteries; atherectomy procedures (8.19%); stent insertions (4.58%); and thrombolysis (6.94%). The difference between individual therapeutic interventions was not statistically significant (P > .05), with the exception of coronary angioplasty, which was lower (P < .01). The incidence of hematoma was not significantly different (P = .163) between diagnostic venography (0.60%) and IVC filter insertions (0.95%). CONCLUSION: This large multicenter studyestablishes current incidences of all procedure-related hematomas. As expected, performing therapeutic arterial procedures increases the incidence of bleeding complications. However, the type of therapeutic vascular procedure is not a major determining factor.

104 • 1:45 PM Efficacy of CT in the Diagnosis of Blunt Aortic Injury

D.A. Zuckerman, MD, St Louis, MO R.M. Durham, MD • W. Luechtefeld, RN D.J. Herr, RN • M.K. Wolverson, MD E. Heiberg, MD, et al PURPOSE: The effectiveness of chest CT in the diagnosis of thoracic aortic injury was evaluated with use of conventional aortography for comparison. MATERIALS AND METHODS: One hundred fiftyfive consecutive hemodynamically stable patients with blunt chest trauma underwent both CT and aortography. Dynamic CT scans were interpreted by four staff radiologists experienced in CT without knowledge of the results of the aortogram or the clinical outcome of the patient. Anyone of 10 CT signs of aortic injury was accepted as a positive CT diagnosis, requiring a confirmatoryaortogram. RESULTS: Among 14 positive aortograms, there was injury to the thoracic aorta (n = 9), subclavian artery (n = 2), and vertebral artery (n = 1). There was also one false-positive aortogram and one ascending aneurysm. In three ofthe eight patients with surgically repaired aortic transection, the CT interpretation varied among the four radiologists. Overall sensitivity of CT scanning was 85%, with a specificity of 53% and an accuracy of 11%. Sensitivity, specificity, and accuracy were not significantly different from those found at routine chest radiography. CONCLUSION: The sensitivity of CT for indicating the need for thoracic aortography is observer dependent. CT is not reliable in excluding aortic injury. 105

• 2:00 PM

Current Utilization of Lower Extremity Venous US for Diagnosis of Deep Vein Thrombosis

E.A. Drucker, MD, Boston, MA • J.A. Kaufman, MD B.B. Faitelson, MD • E.H. Kim, MD T. Kinney, MD • P. Smith, MD PURPOSE: We compared current utilization ofvenous US of the lower extremity for diagnosis of deep vein thrombosis (DVT) with previous use of contrast venography (CV). MATERIALS AND METHODS: We prospectively and consecutively analyzed all weekday lower extremity venous US examination performed in our department over a 4-month period. CV performed during a year prior to the introduction of routine venous US (1985) was also analyzed. RESULTS: A total of 327 venous US examinations were performed during the 4-month study period. Initial analysis of the venous US studies demonstrates the following: positive 5%, both legs scanned 53%, calf scanned 5%. Services ordering venous US studies

1994 SCVIR Meeting Abstracts • 39 Volume 5 Number 1

were orthopedics 53% (72% bilateral), medicine 32% (32% bilateral), general surgery 5%, emergency ward 5%, rehabilitation 5%, and urology 3%. None of the venous US studies were followed by CV. Initial analysis of CV demonstrates the following: positive 28%, bilateral venograms 28%. Services ordering CV were orthopedics 56% (50% bilateral), medicine 16%, and general surgery 11%. CONCLUSION: The substitution of venous US for CV in diagnosis of lower extremity DVT has resulted in an almost sixfold decrease in the rate of positive examinations at our institution. Aside from a doubling of referrals from internal medicine, the percentage of studies ordered by specialty is unchanged. There has been a 50% increase in the number of bilateral leg examinations ordered. These results suggest a lowering of the dinical threshold for ordering venous US to diagnoseDVT. 106

2:15 PM

Ergonomic Vascular Access Needle with Blood-Containment Capability R.J. Gray, MD*, Washington, DC B.L. Dolmatch, MD • J.H. Rundback, MD K.M. Horton, MD PURPOSE: Standard vascular access needles allow uncontrolled spraying of blood through the needle hub. The ergonomic vascular access needle (EVAN) is designed to contain blood while allowing observation of blood pulsations prior to passage of a guide wire. We evaluated the EVAN in 110 patients. MATERIALS AND METHODS: One hundred ten arterial access procedures (93 diagnostic, 17 interventions) were performed with use of the EVAN (Arrow) for initial puncture. In each case, visualization ofpulsatile blood motion within the needle, containment of blood, and guide-wire passage were evaluated. The number of attempts, guide-wire type, access vessel arteriogram, and complications were recorded. RESULTS: A total of92% procedures (101 of 110) were successfully completed, usually within two attempts (89%). Success was independent of guide-wire type. In 90% (99 of 110), contained pulsatile blood motion was observed prior to attempted guide-wire passage. Uncontrolled spraying of pulsatile blood did not occur. In 18% (20 of 110), slow oozing of blood occurred through the needle hub; most (17 of20) of these were during guide-wire passage. Oozing was common with floppy-tipped guide wires (57%) and occurred occasionally with stiffer-tipped guide wires (6%). There were no needle-related complications. CONCLUSION: The EVAN is safe and usually allows successful guide-wire passage into the vascular system. The EVAN eliminates uncontrolled ejection of blood; however, slow periwire leakage occurs frequently if used with a floppy-tipped guide wire.



Physical Examination versus Color Flow Doppler US for Detecting Impending Vascular Graft Failure S.O. Trerotola, MD, Indianapolis, IN P.J. Scheel, MD • N. Powe, MD PURPOSE: Duplex US, specifically volume flow, has been shown to predict failure in hemodialysis access grafts and has been advocated as a routine screening test in all patients with synthetic hemodialysis grafts. This study compared physical examination of the graft with US findings. MATERIALS AND METHODS: Thirty-nine patients were clinically examined, and the results were correlated with subsequent US findings. Graft examination was undertaken by four independent blinded observers who palpated the arterial, midpoint, and venous end of the graft. Patients were categorized as having a palpable thrill at all three locations (group 1), a palpable pulse at any location (group 2), or indeterminate (group 3). RESULTS: The average percentage agreement and kappa scores for the observers were 72% and 55, respectively. Average volume flow was highest for group 1 (1,200 mL/min) and lowest for group 2 (634 mL/ min). Of patients with a volume flow less than 450 mL/min, 100% were classified in groups 2 and 3. Of patients classified in group 1, 100% had a volume flow of greater than 450 mL/min. CONCLUSION: With the acceptable reliability of physical examination between observers, physical examination is a sensitive screening test for determining the need for further investigation of impending graft failure and may obviate US in a significant number of patients. 108

2:45 PM

Thrombotic Events in Young Patients with Hypercoagulable State Due to Elevated Antiphospholipid Antibody Levels J.M. Provenzale, MD, Durham, NC R.C. Gilkeson, MD • P.V. Suhocki, MD B.G. Macik, MD PURPOSE: We assessed the spectrum of thrombotic events in young patients with a hypercoagulable state due to elevated lupus anticoagulant or anticardiolipin antibody (antiphospholipid antibodies) titers. MATERIALS AND METHODS: A retrospective review was performed of patients aged 45 years or younger who had abnormal lupus anticoagulant panels (224 patients) and/ or elevated anticardiolipin antibody titers (52 patients) at Duke University Medical Center during the period January 1992 to August 1993. Patients with a diagnosis of another coagulopathy or systemic lupus erythematosus were excluded.

40 • Journal of Vascular and Interventional Radiology January-February 1994

RESULTS: Twenty patients with thrombotic events (mean age, 31 years) were identified. Six patients had peripheral arterial or venous thromboses, four had mesenteric thromboses, three had pulmonary embolism, 10 had cerebrovascular events, and five had myocardial infarction or severe coronary artery disease. Seven patients had multiple vascular bed involvement, including cerebrovascular accidents and peripheral or mesenteric thromboses (n = 4), coronary artery disease and peripheral arterial thromboses (n = 3), and deep venous thrombosis and pulmonary embolism (n = 2). CONCLUSION: A spectrum of arterial and/ or venous thromboses can be seen in patients with elevated antiphospholipid antibody titers. The presence of otherwise unexplained thrombosis in a young patient should raise the possibility of a hypercoagulable state due to the presence of these antibodies.

Wednesday Mternoon· 6D • Scientific Session 19 Embolization MODERATOR:

Alan M. Zuckerman, MD, Atlanta, GA

109 • 1:30 PM Long-term Follow-up of Percutaneous Embolization of Varicoceles R.D. Shlansky-Goldberg, MD, Philadelphia, PA J.M. Henry, MD • C. Cope, MD • M.C. Soulen, MD z.J. Haskal, MD • K.N. VanArsdalen, MD PURPOSE: Varicocele is the most common treatable cause of male infertility. We present the long-term follow-up of patients treated with percutaneous embolization. MATERIALS AND METHODS: Forty men (25-48 years old) underwent percutaneous embolization of 54 varicoceles from 1988 to 1992. Varicoceles and incompetent gonadal veins on the nonclinically evident side were embolized with Gianturco coils. Semen analysis was performed before and 6 months after treatment. Recurrence, procedure satisfaction, and pregnancy rates were obtained by means of phone interviews, with a mean follow-up of 2.7 years. RESULTS: Mter treatment, semen volume increased 180% (SE, 0.56), sperm density increased 125% (SE, 0.56), and motility improved 26% (SE, 0.50). Phone follow-up revealed no recurrence ofvaricoceles that were clinically evident before embolization. There was one technical failure, for a total operative failure and recurrence rate of 2%. One patient experienced scrotal pain for 4 weeks after embolization (2.5%). Twenty-

one couples (51 %) were able to initiate a pregnancy in the mean follow-up period of 2. 7 years. CONCLUSION: Transvenous embolization of varicoceles is an effective form of primary therapy for clinically significant varicoceles for the treatment of infertility. 110 • 1:45 PM Pelvic Pain Syndrome Treated with Ovarian Vein Embolization R.D. Edwards, FRCR, Liverpool, England P.C. Rowlands, FRCR • J.G. Moss, FRCR I.R. Robertson, FRCR PURPOSE: Pelvic pain syndrome (PPS), without an apparent cause, is a common gynecologic complaint, which presents difficult diagnostic and therapeutic problems. An association between pelvic varices and PPS is well known, but no single effective treatment exists. It has been suggested that the primary problem is retrograde flow in incompetent ovarian veins, similar to the male varicocele. Transcatheter ovarian vein embolization was performed in 10 cases to assess its therapeutic potential in PPS. MATERIALS AND METHODS: Ten female patients with typical clinical features of PPS and normallaparoscopic findings underwent transcatheter embolization of the ovarian veins via the femoral venous route (nine bilateral, one unilateral). RESULTS: Eight patients were cured of their symptoms, one patient reported significant improvement, and one patient did not benefit (mean follow-up, 7.8 months). No adverse effect on ovarian function occurred. Improvement in menstrual symptoms also occurred in six of seven patients. No significant procedural complications occurred. CONCLUSION: Ovarian vein embolization appears to be an effective treatment for PPS associated with pelvic varices. 111 • 2:00 PM Histopathology of Chronic Arterial Occlusions with Hydrogel and Polyvinyl Alcohol Embolic Material in Domestic Swine D.P. Link, Jr., MD*, Sacramento, CA F.A. Mourtada, BSE • R. Virmani, MD J.D. Strandberg, MD • K. Blashka, PhD* PURPOSE: Hydrogel technology offers a wide range of promising possibilities as embolic agents for the future as well as delivery mechanisms for therapeutic agents. There are no studies of the long-term effects. This study compares the effects of hydrogel and polyvinyl alcohol (PVA) emboli under similar conditions. MATERIALS AND METHODS: Seventeen kidneys and five livers were embolized in 14 crossbred domestic swine. Hydrogel emboli were made from polyacry-

1994 SCVIR Meeting Abstracts • 41 Volume 5 Number 1

lonitrile copolymer (PAN), which was used plain and with tantalum loading. Angiographic, gross pathologic, microscopic, and laboratory data were obtained from intervals of 1-50 days. RESULTS: PYA and hydrogel produced similar infarctions, as seen angiographically and on gross pathologic examination. On microscopy, however, the hydrogel emboli produced a more architectural disruption of the embolized arteries. This was characterized in part by a reorganization of the arterial wall smooth muscle cells and disruption of the vessel intima along with the internal elastic membrane. The changes were most pronounced at 4 weeks and have mostly organized at 8 weeks. CONCLUSION: The destructive changes produced by both hydrogel and PVA are surprising and very interesting. The possible mechanisms and implications of this arterial wall response will be discussed.

112 • 2:15 PM Interlocking Detachable Platinum Coils: Early Clinical Experience with This Controlled Embolization Device J.F. Reidy, London, England PURPOSE: The authors present the first clinical evaluation of interlocking detachable coils (IDC) (Target Therapeutics). MATERIALS AND METHODS: The IDC is similar to electrolytic coils (GDC) except that the IDC is mechanically attached to its own pusher wire. Embolization is reversible until this point is beyond the tip of Tracker 18 catheter. Coils used were 0.018 inch with no attached fibers. Range of diameters was 2-8 mm and lengths were 1-20 cm. Thirty-two IDCs were used for eight embolization procedures in six patients for two coronary arteriovenous fistulas (patient ages, 2 and 3 years), one Blalock-Taussig shunt (11 years), one large renal artery aneurysm (60 years), four large aortopulmonary collaterals in pulmonary atresia (13 months), and an innominate arteriovenous fistula (56 years). RESULTS: All coils passed easily (even long coils in tortuous arteries) and were reversible until detachment. Occlusion was achieved in all with no complications. One coil (diameter too small) passed through a Blalock-Taussig shunt but was easily retrieved. CONCLUSION: IDCs represent a technical advance that allow$ a localized mass of platinum wire to be accurately positioned in any shape vascular lesion and to effect precision occlusion of an abnormal artery or fistula.

113 • 2:30 PM Efficacy of Transcatheter Embolization for Intractable Hemorrhage from Bladder or Uterine Neoplasms: Long-term Follow-up E.K. Lang, MD, New Orleans, LA PURPOSE: Control of bleeding from pelvic neoplasms with transcatheter embolization (TCE) is studied for different embolic materials and techniques. MATERIALS AND METHODS: Fifty-three patients underwent embolization for intractable hemorrhage from bladder or uterine neoplasms from 1969 through 1988. Forty-four patients had cancer of the cervix, nine had bladder cancer (five men, four women). IvaIon was used as embolic material in six patients, Gelfoam in 27, cyanoacrylate in nine, autologous blood clot (sometimes re-enforced with amicar) in eight, and wire coils in three. Embolization of both anterior divisions of the hypogastric artery was carried out in 44 patients and was unilateral in nine. If specific bleeding sites were identified, embolization was limited to these branches. RESULTS: TCE was technically successful in 52 patients. Initial control of bleeding was achieved in 42 patients; 10 died of uncontrolled bleeding within the first week. Nine were lost to follow-up (eight with initial control of bleeding). Forty-four patients were followed up from 1 to 180 months. Initial control of bleeding was accomplished with Gelfoam in 20, cyanoacrylate in eight, autologous blood clot in seven, IvaIon in four, and coils in three. Long-term control of bleeding was achieved with Gelfoam in eight, cyanoacrylate in three, autologous blood clot in two, Ivalon in one, and wire coils in one. Four patients survived more than 4 years; eight patients survived more than 2 years. Mean survival is 11 months; maximum survival is 15 years. CONCLUSION: Initial results were best in patients with an identified unilateral bleeding site in whom cyanoacrylate (six of six) or wire coils (three of three) were used as embolic material. Long-term results were best for cyanoacrylate (three of six) and wire coils (one of three). Initial control of bleeding was achieved in 79% with long-term control in 28%. Complications included two perineal sloughs (one with cyanoacrylate, one with Ivalon), one rectal slough (IvaIon), and one bladder contracture (Ivalon). Amicarenforced autologous blood clot and Gelfoam cubes offered the advantage of easy availability, absence of complications, and comparable rates of short- and long-term control of bleeding.

42 • Journal of Vascular and Interventional Radiology January-February 1994

114 • 2:45 PM Percutaneous Occlusion of Patent Ductus Arteriosus: Experimental Study A. Sakinis, MD, Minneapolis, MN J. Cervera-Ceballos, MD • M. Urness C.H Pozza, MD • J.L. Titus, MD • K. Amplatz, MD PURPOSE: We tested the efficacy of a new occlusion device that could be used in the pediatric age group. MATERIALS AND METHODS: In 10 dogs, shunts were made between the descending aorta and pulmonary artery with use of the carotid artery or Goretex (4-6 mm). The shunts were catheterized from the venous side with a 6-F catheter, and the occlusion device was introduced. The device consists of a fine nitinol wire mesh with a bell-shaped configuration. The device is introduced through the catheter into the aorta and is pulled back until it gets stuck, occluding the mouth of the graft. Blood gas measurements were obtained prior to and after plug placement, and a follow-up aortogram and pulmonary arteriogram were obtained one-half hour and 7 days after the procedure. After 3 months the animals were killed. RESULTS: Complete occlusion of the graft occurred in all animals in 40 minutes to 1 month. There were no misplacements, embolizations, or evidence ofpulmonary embolism. CONCLUSION: The new device is highly promising and applicable for clinical use.

Wednesday Mternoon • 6E • Scientific Session 20 Abscess Drainage MODERATOR:

Richard B. Towbin, MD, Pittsburgh,


115 • 1:30PM Outcome Analysis after Percutaneous Drainage of Periappendiceal Abscesses R.M. 0 'Laoide, MD, San Diego, CA H D'Agostino, MD • E. vanSonnenberg, MD C. Pittman, MD • J. Donaldson, MD • E. Hart, MD PURPOSE: The long-term outcome is reported for 17 patients who underwent percutaneous drainage of periappendiceal abscesses. MATERIALS AND METHODS: Seventeen patients (14 men and three women), aged 13-74 years, underwent percutaneous drainage of periappendiceal abscesses. The procedure was performed under CT guidance in 14 patients and US guidance in three patients.

RESULTS: A fistula was identified in seven patients, which closed spontaneously in all patients. Interval appendectomy was performed in nine patients 3-20 weeks after catheter removal (mean interval, 12 weeks). At surgery, adhesions and an appendiceal stump were identified in seven patients; however, there was no evidence of a residual abscess. Pathologic examination of the appendix did not reveal evidence of perforation in any case; however, residual serosal periappendiceal inflammation was frequently seen. Eight patients did not undergo interval surgery, and all have remained asymptomatic (mean follow-up, 27 months). CONCLUSION: Interval appendectomy may not be necessary in all patients after percutaneous drainage of periappendiceal abscesses. The most frequent findings at interval appendectomy were adhesions and no appendiceal stump. 116 • 1:45 PM Image-guided Transvaginal and Transrectal Aspiration and Drainage of Pelvic Collections: Technical Details and Results over a 3-year Period

A.M. Beale, MD, San Diego, CA HB. D'Agostino, MD • B. Stainken, MD G. Casola, MD • E. vanSonnenberg, MD PURPOSE: Routine pelvic imaging often includes transvaginal (TV) or transrectal (TR) endosonography. Therefore, TV and TR interventional procedures are a natural extension of these diagnostic techniques. We present our experience with sonographically guided TV and TR aspiration and drainage for pelvic collections. MATERIALS AND METHODS: From July 1990 to September 1993, 25 procedures (22 TV and three TR) were performed. The origins of the collections included tubo-ovarian abscess (16 patients), infected hematoma (two patients), periappendiceal abscess (three patients), postoperative (three patients), and endometrioma (one patient). Twenty-one collections were infected and four were sterile. Patients were imaged with TV US probes with use of a needle guidance device (23 patients) or a Teflon sheath (two patients) for guidance. Patients with small collections underwent aspiration through an 18-gauge needle (eight TV patients, one TR patient). Single lumen 7-12-F catheters were used for drainage (16 patients). Successful drainage was defined as resolution of the collection by clinical or radiologic parameters. RESULTS: Success was achieved in 23 of 25 patients. Duration of catheterization ranged from 3 to 20 days (mean, 6 days). One patient with a tubo-ovarian abscess required surgical drainage, and the patient with endometrioma required repeated drainage after aspiration.

1994 SCVIR Meeting Abstracts • 43 Volume 5 Number 1

CONCLUSION: Image-guided TV and TR drainages are effective, expeditious, and safe techniques for management of pelvic collections. Patient tolerance to the procedure and drainage catheter is excellent. TV and TR routes are preferred over the transgluteal approach for aspiration or drainage of pelvic collections. 117 • 2:00 PM Transrectal US Guidance for Drainage of Deep Pelvic Abscesses D.J. Eschelman, MD, Philadelphia, PA AA Alexander, MD • D.M. Hovsepian, MD L. Nazarian, MD • J. Bonn, MD PURPOSE: As an alternative to traditional surgical transrectal drainage, we have examined the safety and efficacy of transrectal US-guided drainage for deep pelvic abscess not accessible by percutaneous transabdominal, transgluteal, or transvaginal routes. MATERIALS AND METHODS: In eight febrile patients (four male, four female) ranging in age from 7 to 51 years, cross-sectional imaging (US or CT) revealed pelvic abscesses that were deemed unapproachable by alternate nonoperative routes secondary to interposed bowel (n = 5), presacral location (n = 1), or age (n = 2). A 7.5-MHz endfire transrectal probe (Acoustic Imaging or Advanced Technologies Laboratories) fitted with a biopsy guide was inserted into the rectum and each collection localized. The procedures were performed with use of intravenous sedation (adults) or general anesthesia (children). Under transrectal US guidance an I8-gauge needle was advanced into each collection, and then an 8-I4-F locking drainage catheter was placed fluoroscopically by means of the Seldinger technique in each patient. RESULTS: All collections were successfully accessed and effectively drained without complication. Catheters were removed in 2-10 days without recurrence of abscess. CONCLUSION: Transrectal US-guided drainage of deep pelvic abscesses provides a safe, well-tolerated, effective alternative to the more traditional surgical transrectal drainage, especially in pelvic abscesses lacking a safe percutaneous approach. 118 • 2:15 PM Combined Transrectal Sonographic and Fluoroscopic Guidance for Deep Pelvic Abscess Drainage D.J. Kastan, MD, Detroit, MI • P.C. Shetty, MD M. W. Burke, MD • R.P. Sharma, MD M.A. Angileri, MD PURPOSE: This study was performed to determine the feasibility, safety, success, and patient tolerance for transrectal catheter drainage of pelvic abscesses

with use of a combination of transrectal sonography and fluoroscopic guidance. MATERIALS AND METHODS: Transrectal drainage was performed in five patients (four male, one female) over a 20-month period with use of transrectal sonography (all cases) and fluoroscopic guidance (three cases). The abscesses could not be palpated at rectal examination. Locking-loop self-retaining catheters were placed for drainage. Transabdominal drainage was precluded due to intervening structures that could not be traversed. Transrectal drainage was selected over posterior drainage through the sciatic notch. RESULTS: All five procedures were technically and clinically successful. Clinical improvement was prompt, and all catheters were removed within 7 days. No patients required further intervention, and no complications were detected in our series. CONCLUSION: Transrectal sonographically guided catheter drainage is safe, effective, and surprisingly well tolerated by patients. The use of the Seldinger technique is aided by concomitant fluoroscopy, but this is not absolutely necessary. This method is preferred over the posterior approach (sciatic notch) when the anterior abdominal approach is contraindicated. 119 • 2:30 PM Drainage Efficiency of Double-Lumen "Sump" Catheters and Single-Lumen Catheters: An in Vitro Comparison AC. Hoyt, MD, San Diego, CA • H. D'Agostino, MD AJ. Carillo, MD • S. Kim, MD R.M. O'Laoide, MD • E. vanSonnenberg, MD PURPOSE: The purpose of our study is to compare the drainage efficiency of double-lumen and singlelumen catheters in an in vitro model. MATERIALS AND METHODS: Four I2-F 27.5-cm catheters were used in the study. Three of these had a double lumen and one had a single lumen. The "sump" hole opened internally to the catheter lumen (one catheter) or externally (one catheter), or the "sump" lumen was occluded (one catheter). Catheters were connected to continuous low wall suction and were immersed in water and fluid with particles. The volume of fluid drained by each catheter was measured and compared. RESULTS: The single-lumen catheter drained the largest volume for each type of fluid (203 mL/min with water, and 112 mL/min with particulate fluid). The double-lumen catheter with the "sump" occluded drained 192 mL/min of water and 87 mL/min of particulate fluid. Both double-lumen catheters drained 60 mL/min of water and 77 mL/min of particulate fluid. The open sump allowed air to be suctioned into the

44 • Journal of Vascular and Interventional Radiology January-February 1994

catheter, decreasing the catheter's drainage efficiency in our experimental model. CONCLUSION: Double-lumen catheters performed equally as well as or less efficiently than single-lumen catheters in our in vitro model. A clinical study comparing the efficiency of double-lumen and single-lumen catheters for percutaneous drainage of fluid collections is suggested. 120 • 2:45 PM Multicenter Comparison Study of the Drainage Efficiency of Single- and Double-Lumen Drainage Catheters

H.R. D'Agostino, MD, San Diego, CA R. Niola, MD • L. Romano, MD • G. Rossi, MD R. Villaueiran, MD • B.F. Stainken, MD PURPOSE: We compared the drainage efficiency of single- and double-lumen catheters used for drainage of abdominal fluid collections. MATERIALS AND METHODS: From October 1991 to September 1993, 278 patients underwent percutaneous drainage of abdominal collections at three institutions. Origins of the collections included postoperative (227 patients), after bowel perforation (23 patients), hematogenous (19 patients), and by direct extension (eight patients). Patients who had their collections drained with single-lumen catheters (n = 142) and those who underwent drainage with double-lumen catheters (n = 136) had comparable collections. Catheters were removed when drainage output decreased to no measurable amount. The criterion for successful drainage was resolution of the collection by clinical or radiologic criteria. RESULTS: Single-lumen catheters enabled successful drainage in 127 of 142 patients (89%), and doublelumen catheters in 118 of 136 patients (87%). In both groups failure was attributed to characteristics of the abscess (fistula, necrotic tumor). These differences were not statistically significant. CONCLUSION: The results from this series suggest there are no differences in the drainage efficiency of single- versus double-lumen drainage catheters. The results of this study provide clinical data to assist in catheter selection for percutaneous drainage.

Wednesday Mternoon • 6C • Scientific Session 21 Thrombolysis MODERATOR:

Anne C. Roberts, MD, San Diego, CA

121 • 3:45 PM Long-term Value of Local Thrombolytic Therapy versus Heparin for Treatment of Deep Vein Thrombosis

P.E. Thorpe, MD, Omaha, NE • J.D. Edwards, MD • P.S. Dougan, MD PURPOSE: We present results from a series of 60 patients treated with local urokinase and/ or heparin between 1988 and 1992 to compare the safety and efficacy of treatment and determine if aggressive lytic therapy for deep vein thrombosis (DVT) results in better long-term clinical status. MATERIALS AND METHODS: Thirty of60 patients with DVT-classified as acute in 14 (47%), subacute in 12 (40%), and chronic in four (13%)-received urokinase. Dorsal vein infusion, with intermittent malleolar tourniquet application, was used for legs; multiple side-hole catheters were placed in arms. All patients were given heparin and discharged receiving Coumadin. Interval extremity circumference was recorded. Initial and completion venography was performed in all patients given urokinase. Clinical assessment was performed by means of single-blind telephone interview in 54 of 60 for 90% participation in follow-up ranging from 12 to 65 months. RESULTS: There were five (17%) bleeding complications requiring cessation of therapy, but thrombolysis was achieved in 29 of 30 patients. Average urokinase dose in legs versus arms was 3,220,000 IV versus 1,891,300 IV in arms infused over a mean of 46.3 vs 19.7 hours, respectively. Heparin infusion averaged 154 hours. In comparing results for patients treated with urokinase versus heparin, follow-up revealed (a) no signs of post-thrombotic syndrome, 80.6% versus 31.2%; (b) occasional edema or pain, 19% versus 69%; (c) severe edema, 4% versus 16%; (d) recurrent DVT, 6% versus 30%; and (e) pulmonary embolus, 0% versus 10%. CONCLUSION: Local lytic therapy for DVT gives faster, immediate relief of signs and symptoms of thrombosis. Patients with subacute/chronic symptoms also benefit over the long-term from urokinase. Direct pedal infusion may be more effective for treatment ofinfrainguinal thrombosis than systemic infusion. Long-term follow-up suggests a significant decrease in the incidence of post-thrombotic syndrome after urokinase compared with that in patients with

x.x. Zhan, MD

1994 SCVIR Meeting Abstracts • 45 Volume 5 Number 1

DVT treated with heparin alone. These data suggest the need for a multicenter national randomized study. 122 • 4:00 PM Aggressive Catheter-directed Thrombolysis of Iliofemoral Venous Occlusion M.D. Dake, MD, Stanford, CA • C.P. Semba, MD PURPOSE: Standard therapy for treatment ofiliofemoral venous thrombosis has been heparin and Coumadin. The purpose of this study was to evaluate regional urokinase therapy in the treatment of symptomatic acute and chronic iliofemoral venous thrombosis. MATERIALS AND METHODS: Twenty-five consecutive symptomatic patients with iliofemoral venous thrombosis (acute in 15, chronic in 10; mean, 14.4 months) underwent regional thrombolysis with urokinase. In 15 patients, transcatheter urokinase therapy was initiated after intravenous heparin therapy. Average urokinase dose was 3.6 million U infused over a mean period of27 hours (range, 8-74 hours). Fibrinogen levels were maintained above 100 mg/dL in all patients. RESULTS: Complete lysis occurred in 18 of 25 patients (72%). Sixteen patients had underlying stenoses, which were treated with angioplasty (n = 2) or angioplasty and stent placement (n = 14). One patient has no significant thrombolysis but underwent successful iliac vein recanalization with angioplasty and stent placement. All of 24 patients had rapid resolution of their symptoms after intervention. Twenty-two patients had improvement in symptoms following urokinase alone. One minor complication occurred in a patient who developed heme-positive stools with a drop in hematocrit. Urokinase was stopped, and no transfusion was required. CONCLUSION: This initial experience suggests that regional thrombolysis with urokinase for the treatment of symptomatic iliofemoral venous disease is effective and safe. 123 • 4:15 PM Interventional Management of Proximal Upper Extremity Venous Thrombosis: An 8-year Retrospective Study M.R. Gomes, MD, Minneapolis, MN C.H. Pozza, MD • G.K. Nazarian, MD D.J. Stackhouse, MD • H. Bjarnason, MD D.W. Hunter, MD PURPOSE: We evaluated the interventional management of upper extremity-to-SVC thrombosis and determined immediate and long-term outcomes compared with historical data. MATERIALS AND METHODS: An 8-year retrospective review was performed on all patients who had a

diagnosis of upper extremity venous thrombosis. Ninety patients were identified who received radiologic treatment. Fifty-seven records with adequate follow-up were available for analysis. Forty-two patients had catheter-induced thrombus and 14 had effort thrombosis. Left-sided veins were involved in 44 of 57 patients, although the right side was involved in 10 of 14 patients with effort thrombosis. A combination of peripheral upper extremity and femoral vein approaches to thrombolysis, angioplasty, stent placement, and thrombectomy were used. A variety of thrombolytic regimens were employed. RESULTS: Partial or complete success in eliminating thrombotic occlusion and reestablishing flow was achieved in 85% of patients. The single most important factor determining success was duration of occlusion less than 2 weeks. CONCLUSION: Thrombolytic treatment with adjuvant therapy is effective in re-establishing blood flow in upper extremity venous thrombosis ofless than 2 weeks duration. This results in a rapid return of the patient to an asymptomatic state. 124 • 4:30 PM Percutaneous Local Intraarterial Thrombolysis: Analysis of Factors Affecting Outcome K.R. Stokes, MD, Boston, MA • M.E. Clouse, MD PURPOSE: We evaluated the 30-day patency rate of vein graft and native artery urokinase infusion and analyzed factors affecting outcome. MATERIALS AND METHODS: Eighty-two patients received urokinase infusion into native arteries or vein grafts, and the 30-day patency rate, mortality rate, and factors affecting outcome were evaluated. RESULTS: Thirty-day patency rate was 76.8%, with average infusion of 27 hours. Stenoses in native arteries were treated with PTA or surgery. Graft stenoses were treated operatively. The presence of at least onevessel runoff was the most important factor affecting thrombolytic outcome (94.7% patency). Age of occlusion up to 3 months and heparin dose did not affect thrombolytic outcome. Length and location of graft and catheter type (single or coaxial) did not affect infusion time or thrombolytic outcome. No significant difference in success was found between native artery and graft occlusion, or between embolic versus thrombotic occlusion. Amputation rate was 18%. The 30-day mortality rate was 6.1%, with a procedural-related mortality of 2.4%. CONCLUSION: Successful lysis depends on patent peripheral runoff. Other factors including the degree to which patients were heparinized do not affect lysis.

46 • Journal of Vascular and Interventional Radiology January-February 1994

125 • 4:45 PM Infrainguinal Bypass Graft Entrapment: Angiographic Diagnosis and Lytic Therapy M.C. Soulen, MD, Philadelphia, PA S.E. Braverman, MD • J.P. Carpenter, MD R.D. Shlansky-Goldberg, MD PURPOSE: Prolonged secondary patency after thrombolysis of failed bypass grafts requires identification and correction of the underlying lesion precipitating graft failure. We report four cases of infrainguinal vein grafts in which failure occurred due to popliteal entrapment. MATERIALS AND METHODS: All patients underwent diagnostic angiography for recurrence ofischemic symptoms 1 month to 4 years after infrainguinal bypass. Three had thrombosed grafts that were successfully thrombolysed with urokinase; one graft had no flow but was not thrombosed. RESULTS: No fixed anatomic lesion could be identified to explain graft occlusion. In each case there was flow in the graft with the knee flexed but not in extension. All grafts had an abnormally medial course in their distal portion. In one case, intraarterial pressure in the distal graft dropped by 70 mm Hg and lost phasic waveform with knee extension. Entrapment of the grafts due to improper tunneling through or medial to the medial head of the gastrocnemius was confirmed surgically and corrected by division of the gastrocnemius. CONCLUSION: Femoral-popliteal graft entrapment should be suspected when no fixed anatomic lesion is found and the distal graft has an unusually medial course. Flexion-extension angiograms are diagnostic. Recognition of this entity permits relatively simple surgical correction, sparing the patient a major revascularization procedure. 126 • 5:00 PM Thrombolysis of Thrombosed Popliteal Aneurysms V.H. Leen, MD, Philadelphia, PA R.D. Shlansky-Goldberg, MD • J.P. Carpenter, MD C. Cope, MD • M.C. Soulen, MD • Z.J. Haskal, MD PURPOSE: Popliteal artery aneurysm (PAA) is the most common peripheral aneurysm. The usual treatment is surgical revascularization, which is difficult due to the chronic clot and embolization ofrunoffvessels. Treatment with a thrombolytic agent is controversial due to the risk of further embolization with chronic thrombus and bleeding from other associated aneurysms. MATERIALS AND METHODS: From 1982 to 1993, 14 thrombosed PAAs were treated with thrombolysis (eight from the University of Pennsylvania; six from the STAR registry). All 14 patients were men between

the ages of 30-81 (average, 58) years. All aneurysms were symptomatic for 1-30 days. One patient was treated with streptokinase, while the rest were treated with urokinase. Percutaneous infusions lasted from 4.5 to 48 hours. In all the cases in which urokinase was administered the infusion rate was 60,000240,000 U Ih. RESULTS: Thrombolytic therapy improved runoff in 13 cases (93%); seven patients went on to bypass surgery and six did not. Distal embolization occurred in one patient who was refractory to further therapy but did not require surgery. One infusion was complicated by renal failure. CONCLUSION: Thrombolysis is a effective treatment for thrombosed PAA; however, distal embolization may occur.

Wednesday Afternoon· 6F • Scientific Session 22 Angioplasty: Visceral and Cerebral MODERATOR:

Arina van Breda, MD, Alexandria, VA

127 • 3:45 PM Experience with Palmaz Stents in Renal Artery Stenoses M. Saeed, MD, La Jolla, CA • R.A. Schatz, MD* H.J. Knowles, Jr, MD • R.B. Dilley, MD E.F. Bernstein, MD PURPOSE: Use of vascular stents was evaluated in the treatment of atherosclerotic renal artery stenoses unresponsive to PTA. MATERIALS AND METHODS: Palmaz stents have been deployed in 65 renal arteries in 50 patients. Mean patient age was 65 years. Coexisting coronary artery disease was present in 60%, diabetes in 26%, and renal insufficiency in 38%. Previous PTA had failed in 48%. RESULTS: All stents were successfully deployed. Mean residual stenosis was 2%, with no pressure gradient. Six-month follow-up results showed angiographic restenosis in 36% (75% of eligible patients underwent angiography) and improved blood pressure control in 75%. Among patients with renal insufficiency, 42% improved, 42% showed no change, and 16% had worsening ofrenal function. CONCLUSION: Stents provide excellent initial resolution of stenoses. Longer term salvage can be achieved in the majority oflesions unsuitable for PTA. However, restenosis will continue to be a problem in this unfavorable group oflesions until adjunctive mea-

1994 SCVIR Meeting Abstracts • 47 Volume 5 Number 1

sures evolve in controlling this complex biological phenomenon.

128 • 4:00 PM Palmaz Balloon-expandable Stents for the Treatment of Ostial Renal Artery Stenoses D. W. Trost, MD, New York, NY • T.A. Sos, MD PURPOSE: We prospectively studied the Palmaz renal artery stent in the treatment of ostial renal artery stenoses. MATERIALS AND METHODS: From May 1989 through June 1993, in 17 procedures, we attempted to place 19 stents into 17 renal arteries in 16 patients. (One patient received a second stent in the same artery during a separate procedure.) Eighteen were placed for ostial stenoses and one for a transplant anastomotic stenosis. Indications for intervention were hypertension in 81% (13 of 16), renal failure in 25% (four of 16), and pulmonary edema in 19% (three of 16). Fifty-three percent (10 of 19) were placed after one or several failed angioplasties; 47% (9 of 19) were placed immediately following a suboptimal initial angioplasty; and 95% (18 of 19) of the stents were successfully deployed. There were three complications: branch renal artery thrombosis (n = 1), percutaneous stent removal following unsuccessful placement in a small accessory artery (n = 1), and minor puncture site hematoma (n = 1). RESULTS: Clinical follow-up revealed blood pressure and renal function improvement in 75% (12 of 16) (range, 1-14 months; mean, 6.7 months). Angiographic follow-up has been obtained in 47% (eight of 17) of procedures at 3-8.5 months (mean, 5.75 months); 44% of stents (four of nine) are widely patent, 56% of stents (five of nine) had greater than 50% restenosis. One patient died prior to angiographic follow-up of unrelated disease. CONCLUSION: Renal artery stent placement is a promising treatment for ostial and anastomotic renal artery stenosis that is unresponsive to conventional balloon angioplasty. 129 • 4:15 PM Balloon Angioplasty for the Treatment of Mesenteric Ischemia E.K. Fitzcharles, BS, Charlottesville, VA A.H. Matsumoto, MD • C.J. Tegtmeyer, MD J.B. Selby, Jr, MD • C.G. Tribble, MD J.F. Angle, MD, et al PURPOSE: The efficacy of PTA was evaluated for the treatment of mesenteric ischemia (MS!). MATERIALS AND METHODS: From 1981 to 1993, 19 patients underwent PTA of 20 visceral arteries (13 superior mesenteric, five celiac, two inferior mesenteric arteries). There were 10 male and nine female

patients who had a mean age of 63 years. Fourteen patients had classic symptoms ofMSI or underwent prophylactic dilation prior to another procedure. Five patients had nonclassic MSI symptoms. Follow-up was obtained on all 19 patients. RESULTS: Fifteen of 19 patients (79%) underwent a technically successful PTA. The four technical failures were all secondary to extrinsic arterial compression (three median arcuate ligament, one occult malignancy). Twelve of 15 patients (80%) had immediate clinical improvement. Treatments in four of five patients with nonclassic MSI symptoms were clinical failures. Of the 12 clinical successes, 10 (83%) are still clinically improved at 4-73 months (mean, 28 months). One of the patients in the late failure group underwent technically and clinically successful repeated PTA two times but was lost to follow-up at 27 months. Significant complications occurred in three of the 19 patients. CONCLUSION: PTA of visceral arteries appears to be an effective treatment for patients with classic symptoms ofMSI.

130 • 4:30 PM

Outcome of Revascularization for Chronic Mesenteric Ischemia: Comparison of PTA and Operative Arterial Bypass Grafting S.C. Rose, MD, Seattle, WA • T.M. Quigley, MD E.J. Raker, MD PURPOSE: The outcomes of mesenteric PTA and operative arterial bypass grafting were compared. MATERIALS AND METHODS: Charts were reviewed on all patients who underwent a revascularization procedure since 1985 for suspected chronic mesenteric ischemia. PTA was performed nine times in eight patients (mean age, 70 years). Arterial grafts were placed in nine patients (mean age, 63 years). RESULTS: Seven of nine PTA procedures were initial angiographic successes and resulted in initial symptom improvement: one recurred within 7 days; two improved patients died of other causes at 10 days (CHF) and 1 year; four continue to be improved from 10 weeks to 18 months. Complications included a dissection of the superior mesenteric artery, a brachial arterial hematoma, and a mesenteric cholesterol embolization syndrome. All nine operative procedures were technically successful and resulted in symptomatic improvement. One patient died at 9 days (ARDS). One improved patient died at 7 months (stroke); the remaining seven continue to be improved from 15 months to 8 years. Complications included a splenic laceration and atheroembolization to the toes. CONCLUSION: Patients with classic symptoms and angiographic findings of chronic mesenteric ischemia who are relatively good operative risks should undergo more durable arterial grafting. PTA is satisfactory for

48 • Journal of Vascular and Interventional Radiology January-February 1994

patients with angiographic stenoses and either atypical symptoms or prohibitive operative risks. 131

• 4:45 PM

Cerebral PTA R. T. Higashida, MD*, San Francisco, CA F.Y. Tsai, MD • V.V. Halbach, MD* C.F. Dowd, MD • G.P. Teitelbaum, MD C.G. McDougall, MD, et al PURPOSE: We evaluated the safety, efficacy, and results of PTA on hemodynamically significant ( > 70%) lesions involving the extra- and intracranial vessels of the head and neck. MATERIALS AND METHODS: Since 1981, 274 patients ranging in age from 9 to 78 years were treated with transluminal angioplasty techniques for highgrade stenotic lesions; 125 cases in the anterior circulation and 149 cases in the posterior circulation were treated. Patients were followed up clinically at 1-3 and 6-12 months after treatment. RESULTS: Successful angioplasty with improvement in luminal diameter (> 60%) was achieved in the majority of cases. Permanent complications related to treatment included seven (2.6%) strokes, of which five were due to distal thromboemboli and two to vessel rupture, and included three deaths (1.1%). Nine patients (3.3%) developed symptoms of transient cerebral ischemia, which resolved within 24 hours with volume expansion, systemic anticoagulation, and/or antispasmodic drugs. Radiologic follow-up from 3 to 12 months demonstrated 15 (5.5%) cases ofrestenoSIS.

CONCLUSION: In patients with clinically significant stenosis involving the extra- and intracranial territories that is refractory to medical therapy, PTA may be of significant benefit in alleviating symptoms and improving blood flow to the cerebral circulation. 132 • 5:00 PM

Intracranial Carotid Artery Angioplasty for Symptomatic Stenosis D.A. Eckard, MD, Kansas City, KS FL. Siegel, MD • S. Batnitzky, MD P.L. O'Boynick, MD • J. Thomas, MD A. Hermreck, MD PURPOSE: PTA was evaluated for treating symptomatic intracranial internal carotid artery (ICA) stenotic lesions not amenable to surgical therapy. MATERIALS AND METHODS: At presentation, five patients had either transient ischemic attacks (TIAs) (three patients) or strokes (two patients) and were found at angiography to have severe stenoses (> 90%) involving the ICA (one petrous segment, two cavernous segment, two supraclinoid segment). Angioplasty was performed with a Stealth single-lumen PTA dila-

tion catheter, and the lesions were dilated to 3.0 or 3.5 mm. The patients were evaluated following dilation with transcranial Doppler and MR imaging or CT. Clinical follow-up ranges from 7 to 16 months. RESULTS: Angiographically, excellent results were achieved in two patients, good results were achieved in two patients, and a suboptimal result was achieved in one patient who showed no change in the appearance of an elastic lesion. There were no neurologic complications. None of the patients have had a stroke since the procedure. The patient with suboptimal dilation experienced additional TIAs following the procedure but is currently improved with aspirin therapy. None of the other patients have experienced any TIAs since the procedure. CONCLUSION: These early results suggest intracranial ICA angioplasty shows promise in treating patients with severe symptomatic intracranial ICA stenOSIS.

Wednesday Mternoon· 6D • Scientific Session 23 Failing Dialysis Access MODERATOR:

David W. Hunter, MD, Minneapolis,

MN 133 • 3:45 PM Venous Stenoses Associated with Polytetrafluoroethylene Dialysis Access Grafts: Anatomic Location and Results of Angioplasty R. Y. Kanterman, MD, St Louis, MO T.M. Vesely, MD • T.K. Pilgram, PhD • B. W. Guy D. W. Windus, MD • D. Picus, MD, et al PURPOSE: Our objective was to determine the distribution of stenoses and results of angioplasty in patients with identical polytetrafluoroethylene (PTFE) forearm dialysis grafts. MATERIALS AND METHODS: We reviewed 350 consecutive images of dialysis fistulas obtained from January 1, 1990, to June 30, 1992, yielding 90 patients with identical, unrevised PTFE forearm dialysis grafts. We analyzed the location, morphology, and results of angioplasty for each stenosis. Surgical, radiologic, and dialysis records were reviewed to document the subsequent function of each patient's access. RESULTS: On the initial diagnostic images, 101 stenoses were identified. The anatomic distribution included 45% at the venous anastomosis, 12% within 1 em of the anastomosis, 28% in the basilic vein, 7% in the subclavian vein, and 8% in the graft or arterial anastomosis. In our series, 52% of graft malfunctions

1994 SCVIR Meeting Abstracts • 49 Volume 5 Number 1

occurred within the 1st year after graft placement. Life-table analysis revealed a 6-month patency of 63% and a I-year patency of 41% for the first angioplasty in a given graft, and a 6-month patency of 50% and I-year patency of 25% for the second angioplasty. CONCLUSION: Serial venous angioplasties may prolong the life of a graft but with diminishing patency rates over the course of time and multiple interventions. 134


Angioplasty versus Directional Atherectomy of Dialysis Graft Venous Anastomotic Stenoses G.M. Gaylord, MD, Indianapolis, IN T.E. Taber, MD • K.O. Ehrman, MD P.O. Brown, RN PURPOSE: Angioplasty and atherectomy were compared for the treatment of failed dialysis access grafts. MATERIALS AND METHODS: We retrospectively reviewed 230 patient charts and radiographic reports. The Wilcoxon test and Kaplan-Meier analysis were used to determine survival and failure rates. RESULTS: For patients undergoing an initial percutaneous treatment for a failed graft, atherectomy offered statistically significant (P = .045) survival at 30, 180, and 365 days (84.6%, 58.7%, and 40.3% survival) compared with angioplasty (66.6%, 37.5%, and 28.9%) or combined angioplasty/atherectomy (81.8%, 40.8%, and 27.9%). For a second percutaneous procedure, there was no significant difference (P = .84) between the groups. CONCLUSION: For initial treatment of a failed dialysis graft due to venous anastomotic stenosis, atherectomy offers superior survival compared with angioplasty. For the second treatment of a previously treated graft, there is no difference between the procedures. Treatment of an immediately unsuccessful angioplasty with atherectomy will result in survival similar to that for successful angioplasty but not equivalent to a successful atherectomy. 135

4:15 PM

Comparison of Percutaneous Angioplasty, Atherectomy, and Vascular Stent Placement in Failing Hemodialysis Grafts F. Castaneda, MD, Peoria, IL • M.A. Herrera, MD A. Mendez, MD • J.L. Mancera, MD PURPOSE: The purpose of this retrospective study was to evaluate and compare the efficacy of PTA, atherectomy, and stent placement as the sole modality of revascularization of failing hemodialysis graft shunts. MATERIALS AND METHODS: A total of 72 interventions were performed in a 3-year period; of these, 32 were PTAs, 30 were atherectomies, and 10 were

stent placements. Graft thrombosis, the need for surgical revision, and the need for subsequent intervention were used as endpoint events. The period of time from revascularization procedure to the endpoint event was taken as the duration of efficacy. RESULTS: The duration of efficacy at 1,2, 3, 6, 12, 18, 24, 30, and 36 months was 87%, 84%, 78%, 59%, 31%, 18%, 6%, and 0% for PTA; 86%, 80%, 73%, 63%, 50%, 43%, 23%, 3%, and 0% for atherectomy; and 90%, 80%, 70%, 50%, 30%, 20%, 10%, and 0% for stents, respectively. CONCLUSION: If the graft duration of efficacy is greater than 6 months, atherectomy seems to significantly improve shunt patency compared with PTA and stent placement in this study. However, after 30 months, persistent patency is rare, and ultimately further interventions or placement of new access is required. 136

4:30 PM

Pulse-Spray Thrombolysis of Dialysis Access Fistulas: Comparison of Urokinase and tPA H.R. Pagan-Marin, MD, Boston, MA • J. Polak, MD M. Berger, MD • O. Kwon, MD • C. Grassi, MD M. Meyerovitz, MD, et al PURPOSE: Thrombolytic therapy of clotted dialysis access fistulas is an accepted alternative to surgical thrombectomy. Few studies have looked at the cost of the procedure and the efficacy of different agents such as tPA and urokinase. PATIENTS AND METHODS: Twelve patients referred for fistula thrombolysis were compared: six recent patients receiving tPA and six patients who had undergone urokinase thrombolysis. Total thrombolytic dose, time to return of palpable pulse, total procedure time, postprocedural hemostasis time, complications, and cost of agent were monitored. RESULTS: All twelve clotted fistulas were successfully lysed. Average dose was 8.3 mg oftPAand 437,500 U of urokinase. Time to return of pulse (24.2 minutes for tPA, 20.8 minutes for urokinase) and total procedure time (188 minutes for tPA, 189.2 minutes for urokinase) were not significantly different. Time to achieve hemostasis was significantly higher for tPA (59 minutes) than for urokinase (35.8 minutes). Cost was higher for urokinase than for tPA. CONCLUSION: A nonrandomized comparison of urokinase and tPA as lytic agents for pulse-spray thrombolysis of dialysis fistulas showed equal effectiveness and rapidity in restoring flow. Cost analysis tends to favor the use of tPA. A larger randomized trial will be needed to confirm this initial impression.

50 • Journal of Vascular and Interventional Radiology January-February 1994

137 • 4:45 PM Pullback Atherectomy for Hemodialysis Access R.J. Gray, MD, Washington, DC B.L. Dolmatch, MD • K.M. Horton, MD PURPOSE: Balloon angioplasty and directional atherectomy frequently have short-lived results for stenoses associated with hemodialysis. We report results of a phase I trial of the pullback atherectomy catheter (PAC) for hemodialysis access-related stenoses. MATERIALS AND METHODS: Six intragraft and six venous stenoses in nine patients were treated with the PAC (Arrow). Two lesions required adjunctive balloon angioplasty, and two required adjunctive directional atherectomy. Venographic follow-up was obtained for all patients. All specimens were examined histologically. RESULTS: Eighty-three percent of stenoses (10 of 12) were treated with initial technical success. The stenoses were reduced from an average 72% (range, 50%-99%) to 35% (range, 0-60%). For intragraft stenoses, the 6-month primary patency was 60% (three of five) and the 6-month secondary patency was 80% (four of five). All of six venous stenoses restenosed or thrombosed at a mean follow-up of 1 month (range, 1 day to 3 months). All specimens contained fibrous plaque and intimal hyperplasia. In addition, all six venous outflow stenosis specimens contained media. Significant complications included two venous pseudoaneurysms and one PAC catheter-tip fracture. CONCLUSION: On the basis of preliminary results, PAC appears safe and effective for intragraft stenoses; however, it is not safe or effective for venous stenoses. 138 • 5:00 PM Percutaneous Embolization of Dialysis Grafts A.M. Zuckerman, MD, Atlanta, GA L.G. Martin, MD • C.A. Lewis, MD F.S. Sandhu, MD • S.L. Kaufman, MD PURPOSE: Steal from the distal extremity and swelling sometimes complicate dialysis graft placement. Emobilizations of grafts have been performed at our institution to reverse these complications. Although we suspect graft embolization is commonly performed, to our knowledge it has not been reported. This abstract reports our experience. MATERIALS AND METHODS: Five patients have undergone percutaneous occlusion of dialysis grafts at our institution. The patients' charts were reviewed retrospectively. RESULTS: Four grafts were embolized for steal and one was embolized because the graft had been placed proximal to a completely occluded subclavian vein with resultant arm swelling. Four grafts were embolized with Gianturco coils only. One was embolized with coils and gelatin sponge. Of the grafts causing steal,

two thrombosed immediately after embolization, and symptoms in these patients resolved immediately. The other two grafts had no measurable Doppler pulse 24 hours after embolization. Symptoms in one of these patients resolved after 24 hours, and symptoms in the other patients resolved only after occupational therapy. The patient whose graft was embolized for swelling had no edema at physical examination 2 months after occlusion. There were no complications. CONCLUSION: Percutaneous embolization of dialysis grafts is safe, very easy to perform, and effective for reversing complications of dialysis grafts.

Wednesday Mternoon • 6E • Scientific Session 24 Biliary/ Gallbladder MODERATOR:

Robert L. Vogelzang, MD, Chicago, IL

139 • 3:45 PM Percutaneous Cholecystostomy: Experience in 200 Patients G.R. Saboeiro, MD, St Louis, MO • T.M. Vesely, MD D. Picus, MD • M.E. Hicks, MD • M.D. Darcy, MD D.M. Hovsepian, MD PURPOSE: The clinical effectiveness and safety of percutaneous cholecystostomy were evaluated in the treatment of cholecystitis or prior to the percutaneous removal of symptomatic gallstones. MATERIALS AND METHODS: From January 1987 to June 1993, 202 patients underwent percutaneous cholecystostomy tube placement. This group consists of 109 male and 93 female patients, with a mean age of 65.5 years (range, 20-97 years). Seldinger technique was used in 98% of placements and trocar technique in 2%. This group includes 167 patients with acute cholecystitis, five patients with chronic cholecystitis, 24 patients with symptomatic gallstones prior to percutaneous removal, and six patients with fever of unknown origin. Patients were evaluated for clinical response and complications both during and following the procedure. RESULTS: Of those 172 patients with US or nuclear medicine evidence of cholecystitis, approximately 85% showed a significant clinical response within 48 hours following cholecystostomy, as determined by decreased fever, white blood cell count, or pain. There was no significant difference in response between patients with acute calculous versus acalculous cholecystitis. Complications occurred in 7.9% of patients, the majority of which were minor. There were no deaths.

1994 SCVIR Meeting Abstracts • 51 Volume 5 Number 1

CONCLUSION: Percutaneous cholecystostomy tube placement is a safe and effective procedure for the treatment of cholecystitis or as part of percutaneous gallstone removal. 140 • 4:00 PM Percutaneous Cholecystostomy for the Diagnosis and Treatment of Acute Cholecystitis R.L. Vogelzang, MD, Chicago, IL • L.D. La, MD AA Nemcek, Jr, MD • M.A. Braun, MD PURPOSE: We evaluated the efficacy and results of percutaneous cholecystostomy (PC). MATERIALS AND METHODS: Fifty-three patients underwent US-guided PC; all patients were thought to have cholecystitis because of right upper quadrant pain, positive imaging tests, fever, elevated white blood cell count, and/ or elevated liver function tests. Following PC, patients underwent cholecystectomy, eventual removal of the drainage catheter, or longterm drainage. RESULTS: PC was successful in all. Forty-four of 53 patients (83%) had a final diagnosis of acute cholecystitis. All 24 with suspected stone disease, but only 20 of 29 (69%) with suspected acalculous cholecystitis, had cholecystitis. Of the 44 patients, 38 (86%) benefited from PC; twenty-two of24 patients with stones and 16 of20 with acalculous disease benefited. The four patients with acalculous disease who did not benefit had gangrenous cholecystitis. There were seven minor complications and one death due to a lacerated liver. Seventeen of 44 (39%) with cholecystitis underwent cholecystectomy with minimal morbidity, 20 (45%) had their catheters pulled, and seven had longterm drainage. CONCLUSION: PC is safe and effective; it allows not only treatment but diagnosis of cholecystitis. Patients undergo operation under improved circumstances or can have their catheters removed. The incidence of gangrenous cholecystitis, however, indicates that those who fail to respond should undergo urgent surgery. 141

4:15 PM

Heat Ablation of the Normal Gallbladder in Pigs B.D. Aagaard, MD, San Francisco, CA R.L. Gordon, MD • L.A Wetter, MD* C.K. Montgomery, MD PURPOSE: A preliminary animal study investigated the safety and efficacy of ablating gallbladder mucosa with a percutaneously placed heater catheter. MATERIALS AND METHODS: The study was performed in three stages with use of 39 pigs and modification of equipment at each stage. In stage 1, the heater catheter was progressively improved and temperature settings for stage 2 were defined. In stage 2,

the predetermined settings were used in conjunction with mechanical mixing and cystic duct ligation. In stage 3, heat ablation of the cystic duct was added to reduce epithelial regeneration. RESULTS: Thermal ablation at 54°C for 35 minutes was completely successful in 25%, was partially successful in 50%, and failed in 25% of animals. Cystic duct ablation improved overall results and appears vital in removing duct epithelium as a source for mucosal regeneration. Mechanical mixing of the intraluminal contents was essential for even heat distribution for ablation and reduction of adjacent organ damage. Injury to adjacent organs, reversible in the majority of animals, was present when ablation temperatures exceeded 54°C and serosal temperatures of adjacent organs exceeded 43°C for the duration of treatment. CONCLUSION: We conclude from these initial studies that defunctionalization of the gallbladder is potentially achievable with use of thermal techniques, but the thermal range between complete ablation and adjacent organ injury is narrow. 142 • 4:30 PM Prospective Comparison of Two Projection Techniques for the Detection of Intraluminal Signal Voids on MR Cholangiograms and MR Venograms G.A. Holland, MD, Philadelphia, PA • J. Listerud R.A Baum, MD • J. Carpenter, MD H. Y. Kressel, MD • C. Cope, MD PURPOSE: The purpose of this study was to determine if summation of intensities after soft thresholding (SIST) reconstruction techniques can improve on the limited ability of the standard maximum intensity projection reconstruction algorithm (MIP) to detect stones and clots on MR cholangiograms and MR venograms, respectively. MATERIALS AND METHODS: Data from 40 patients were processed with the SIST and MIP algorithms: 20 MR cholangiograms and 20 MR venograms in patients with known disease. All imaging was performed at 1.5 T. SIST and MIP were performed on the same data. MR studies were blindly reviewed by three radiologists. Contrast venograms and cholangiograms served as the gold standard for the MR venographic and MR cholangiographic studies, respectively. RESULTS: SIST reconstructed images demonstrated eight stones and seven clots not seen on MIP reconstructed images. In addition, SIST images better demonstrated the proximal extent of six clots obscured by MIP. There was no discordance between the SIST reconstructed MR venogramsand MR cholangiograms and contrast venograms or retrograde cholangiograms, respectively.

52 • Journal of Vascular and Interventional Radiology January February 1994

CONCLUSION: The SIST technique is superior to MIP technique for identifying clots and stones on projectional MR venographic and MR cholangiographic studies, respectively. If SIST is not available then the source images must be carefully scrutinized, since MIP obscures the lesions. 143 • 4:45 PM Prospective Evaluation of MR Cholangiography G.A. Holland, MD, Philadelphia, PA • T. Meaken R.A. Baum, MD • J. Carpenter, MD M.D. Schnall, MD, • C. Cope, MD PURPOSE: The purpose of this study was to assess the accuracy of MR imaging with FSE MR cholangiography in determining the level and cause of biliary obstruction. MATERIALS AND METHODS: MR cholangiography was performed in 81 patients with a fat-suppressed heavily T2-weighted FSE sequence. Imaging was performed at 1.5 T with the following parameters: 3-mmthick coronal sections, TRITE = 4,000/160, 24-cm field of view, 256 x 256 matrix, with two to four acquisitions. Clinical correlation was obtained from the operative notes, discharge summaries, and pathology reports. Studies were blindly reviewed by three radiologists. Contrast cholangiographic studies or operative findings served as the gold standard in 46 patients. RESULTS: Accuracy ofMR cholangiography in determining level and cause of biliary obstruction was 98% (40 of 41) and 76% (31 of 41), respectively. The accuracy of diagnosis for the various etiologies of obstruction were as follows: pancreatic carcinoma 89% (17 of 19), cholangiocarcinoma 86% (six of seven), pancreatitis 60% (three of five), and stone 83% (five of six). MIP reconstructed images obscured the stones in all cases. TRAP images allowed visualization of four of six stones and five of six stones were identified on the souce images. Normal biliary and pancreatic ducts were visualized in four of five patients. CONCLUSION: FSE MR cholangiography can demonstrate dilated and normal-caliber biliary ducts. In patients with biliary obstruction, MR cholangiography accurately predicts the level and cause of obstruction but individual source images must be carefully scrutinized.

144 • 5:00 PM Cholangioscopic Treatment of Biliary Complications after Liver Transplantation T.K. Roeren, MD, Heidelberg, Germany G.M. Richter, MD, • G. Noeldge, MD G. Otto, MD, • A. Stiehl, MD G. W. Kauffmann, MD PURPOSE: Percutaneous transhepatic cholangioscopy (PTCS) was used as a last therapeutic means in patients with bile duct stenoses and severe necrosis after liver transplantation. The clinical benefit was evaluated. MATERIALS AND METHODS: Nine patients were referred for PTCS 6-20 weeks after liver transplantation. All patients had increasing serum levels ofbilirub~n, AP, and transaminases with severe cholangitis, bIle duct strictures, and additional stones (three patients). All had undergone multiple endoscopic retrograde cholangiopancreatography procedures, and three patients underwent percutaneous transhepatic biliary drainage with therapeutic attempts. These treatments yielded no significant changes in the course of the disease. PTCS was performed up to five times per patient with use of a 10-F fiberscope and various tools to diagnose and treat underlying lesions and to guide balloons or baskets into otherwise unaccessible ducts. RESULTS: PTCS was technically possible in eight of nine patients. Underlying cause for cholangiographic changes in all patients was severe necrosis of bile duct epithelium with resulting stenoses andlor bile encrustation and stone formation. In one patient therapy was not attempted because of severe liver failure and complete necrosis of all inspected bile ducts. Cholangioscopically guided therapy was clinically successful and definitive in five of seven patients (mean followup, 9 months). One patient is still being treated, and in one patient treatment of central bile duct stenoses failed because of additional unaccessible peripheral stenoses. The patients' clinical course correlated well with cholangioscopic findings, while cholangiography was lagging behind. CONCLUSION: Considering the fact that PTCSguided therapy was only indicated in patients in whom other treatment modalities had failed, an overall clinical success rate of about 70% seems promising. Retransplantation as the only therapeutic alternative has a high mortality and was only necessary in two of our patients. PTCS also seems more sensitive to monitor therapeutic efficiency compared with cholangiography. It should be noted, though, that our observation period is still short and that more frequent indication of this expensive modality is not warranted yet.

1994 SCVIR Meeting Abstracts • 53 Volume 5 Number 1

Thursday Afternoon • 6C • Scientific Session 25 Thrombolysis MODERATOR:

Karim Valji, MD, San Diego, CA

145 • 1:30 PM Use of High-Frequency Ultrasound for Thrombolysis R.D. Shlansky-Goldberg, MD, Philadelphia, PA C. Sehgal, PhD • D.C. Redd, MD PURPOSE: We previously demonstrated thrombolysis with use of 20-kHz ultrasound transmitted over a wire. Transverse vibrations, which worsen with bending, limit the ultrasound transmission of this system. Small transducers attached to the end of a catheter would eliminate these pitfalls; however, higher frequencies must be used. We tested a smaller I-MHz 3-mm crystal as a prototype for a transducer that could be attached to a catheter. MATERIALS AND METHODS: Six in vitro experiments were performed with 2-mL aliquots of blood from four donors. Clots were incubated at 37°C for 3 hours and weighed. Three experimental groups were treated for 30 minutes: Clots were sonicated with urokinase (5,000 U per 1 mL of saline) (ultrasound and urokinase group) or with 1 mL of saline only (ultrasound group), in addition to a control group treated with urokinase only (urokinase group). Percentage lysis was determined by weight change and analyzed with an unpaired t test. RESULTS: In the group treated with urokinase and ultrasound, lysis was 15.1% ± 5.1; for ultrasound only, 17.38% ± 7.6; and for urokinase only, 9.9% ± 5.8. There was no significant difference between results with urokinase and ultrasound treatment compared with the ultrasound treatment only. The combined lysis with ultrasound was 16.6% ± 6.4. CONCLUSION: Combining the ultrasound-treated groups, ultrasound thrombolysis was 1.6 times greater than urokinase alone (P < .06). 146 • 1:45 PM In Vivo Treatment of Infected Prosthetic Graft Material with Urokinase: An Animal Model D.A. Nakamoto, MD, Cleveland, OH M.L. Rosenfield, MD • J.R. Haaga, MD K. Merritt, PhD • P.B. Sachs, MD M.C. Hutton, MD, et al PURPOSE: Pyogenic infection of vascular grafts represents a serious complication that may necessitate removal of the infected graft. If better treatment methods could be developed, perhaps some infected

grafts could be salvaged and not removed. This study reports an animal model for evaluating the sterilization of contaminated vascular graft material implants with urokinase and antibiotics. MATERIALS AND METHODS: Polytetrafluoroethylene (Gore-Tex) implants were incubated overnight in a known concentration of bacteria (Staphylococcus epidermidis) and then were implanted subcutaneously into four groups of anesthetized hamsters. The first group (control) received no treatment. The second group received urokinase injections twice daily into each abscess. The third group received intra-abscess urokinase and systemic gentamicin twice daily. The fourth group received only systemic gentamicin. The hamsters were killed after 1 week. The graft implants and surrounding tissues were excised and submitted for quantitative cultures. RESULTS: With use ofa cutoff value of 100 organisms per milliliter, below which the abscesses were considered noninfected, the following rates of noninfectivity were observed: group 1 (control), 5% noninfected; group 2 (urokinase only), 19.4%; group 3 (urokinase and gentamicin), 63.2%; and group 4 (gentamicin only), 32.5%. The noninfectivity rate of group 3 was significantly higher than all other groups combined (P < .001) and was significantly better than that in group 4 alone (P = .013). CONCLUSION: Our results show that the combination of intra-abscess urokinase and systemic gentamicin is very synergistic in graft sterilization. We propose that urokinase may assist in the degradation of both fibrin and the biofilm produced by S epidermidis, thus improving penetration of antibiotics and local host defense mechanisms. 147 • 2:00 PM Development of a Porcine Chronic Clot Model for the Evaluation of Thrombolytic Devices and Agents S.C. Goodwin, MD, Los Angeles, CA M. Rosedale, BS • T.O. McNamara, MD J. Ko, MD • C. Yoon, MD PURPOSE: A porcine chronic clot model was developed to test thrombolytic devices and agents. Currently, no simple chronic clot model is available. MATERIALS AND METHODS: Sixteen domestic swine were used in the study. A stenosis was created in the left superficial femoral artery. One week later 5 mL of autologous blood clot was deposited above the stenosis. The animals were later killed following angiography: four each at 2 days, 10 days, 1 month, and 3 months after clot placement. The thrombosed artery was sent for pathologic evaluation. RESULTS: In 12 of the 16 animals, occlusion of the superficial femoral artery with a large degree of collat-

54 • Journal of Vascular and Interventional Radiology January-February 1994

eralization around the occlusion was demonstrated at follow-up angiography. Four animals showed recanalization. On pathologic examination, animals in the 2-day group showed no thrombus organization. Animals in the 10-day group showed thrombus that was organized only at the periphery. Animals in the I-month group had nearly completely organized thrombus, while animals in the 3-month group had completely organized thrombus. CONCLUSION: This paper outlines a simple method for the creation of a chronic clot model in the domestic swine. One month is required to develop the model. We have used this model to test various thrombolytic agents. 148 • 2:15 PM Effect of Intrathrombic Deposition of a Specific Thrombin Inhibitor (Argatroban) on Pulse-Spray Thrombolysis in an Experimental Clot Model K. Valji, MD, San Diego, CA • J.J. Bookstein, MD* PURPOSE: We evaluated the efficacy of intrathrombic (IT) deposition of the specific thrombin inhibitor Argatroban during pulse-spray thrombolysis (PST) in experimental venous thrombosis. MATERIALS AND METHODS: Clots were produced in the IVC of 53 rabbits by placement of steel coils and balloon injury of the vessel. After 2 days, clots were treated with use of pulse-spray technique. All rabbits underwent PST with 3 mg oft-PA given over 1 hour. One group received adjunctive in IT heparin (500 U). Other groups received intravenous Argatroban (0.30 or 6.0 mg) or IT Argatrobin (0.30 or 3.0 mg) during PST. After treatment, the rabbits were killed and residual clot was weighed. Pretreatment clot weight was estimated from initial clot length with use of a linear regression equation derived from a separate group of untreated animals. RESULTS: PST with t-PA and adjunctive IT heparin caused greater lysis than with t-PA alone (P = .02). Addition of intravenous Argatroban failed to increase the extent of lysis compared with t-PA alone. IT Argatroban at both tested doses enabled further enhancement oflysis compared with t-PA alone or t-PA with IT heparin (P < .02). CONCLUSION: In a rabbit thrombosis model, the use of IT Argatroban during PST with t-PA enabled a significant increase in the extent of clot lysis compared with t-PA alone or t-PA with IT heparin.

149 • 2:30 PM Evaluation of the Dispersal Patterns of Constant Infusion Thrombolysis Catheters in an in Vitro Gelatin Model T.J. Maginot, MD, Minneapolis, MN D. W. Hunter, MD • H. Bjarnason, MD PURPOSE: We evaluated the dispersal patterns of multiple-side-hole infusion catheters/guide wires used in regional thrombolytic therapy with an in vitro gelatin model of vessellgraft thrombosis. MATERIALS AND METHODS: A circumferential clamp placed on one end of gelatin-filled tubing simulated a moderate (50%-70%) or severe (90%-95%) distal stenosis. Sixteen different catheter/guide-wire arrangements were placed in proximal, middle, and distal locations in the tubing. Blue dye was infused at 10 mL/h or 40 mL/h for 15 minutes. The resulting dispersal patterns were observed visually and photographed; 152 infusions were performed. RESULTS: The dye exited through the side holes at the point of least resistance, which varied with catheter/guide-wire position and degree of distal stenosis. End-hole occlusion and higher infusion rate balanced the dispersal pattern among the side holes. The EDM catheter and the Katzen wire had the most consistently homogenous side-hole dispersal patterns. CONCLUSION: In this gelatin model of vessell graft thrombosis, the dispersal of infused dye followed the path ofleast resistance. This may be difficult to predict in clinical situations. The use of end-hole occlusion and higher infusion rates may broaden thrombolysis, thereby decreasing the need for catheter repositioning. 150 • 2:45 PM Procoagulant Effects of Rapid Intravascular Saline Injections J.J. Bookstein, MD, San Diego, CA • K. Arun, MD PURPOSE: Blood cell damage and hypercoagulability were evaluated after non-contrast intraarterial injections. MATERIALS AND METHODS: A 5-F side-slit catheter was passed into the upper abdominal aorta of 13 rabbits, through which 2-mL injections of saline were made at 1,000 psi. A second short end-hole catheter was passed from the opposite femoral artery to the aortic bifurcation for blood sampling during each injection. Activated clotting times (ACTs) and plasma hemoglobin level were determined on control and postinjection samples. RESULTS: Injections of 2 mL of saline through a saline-filled catheter (n = 4) decreased the ACT by 51% ± 24, and raised plasma hemoglobin concentration by 400% ± 404. If the catheter contained blood during saline injection (n = 11), the ACT decreased by

1994 SCVIR Meeting Abstracts • 55 Volume 5 Number 1

61% ± 31, and the plasma hemoglobin level increased 1,660% ± 1,312. Extreme hypercoagulability developed in four rabbits, with ACTs of 0-3 seconds. CONCLUSION: Rapid transcatheter injections of pure or bloody saline can injure red blood cells, presumably activate platelets, and lead to hypercoagulability of a bolus of blood. These results suggest the advisability of anticoagulation for all rapid intravascular injections and the imperative of clearing the catheter of blood before injection.

Thursday Mternoon • 6F • Scientific Session 26 Cancer Diagnosis/Therapy MODERATOR:

Douglas M. Coldwell, MD, PhD,

Seattle,WA 151

1:30 PM

Percutaneous Alcohol Injection for Ablation of Experimental Tumors in an Animal Model M. V. Veira, MD, San Diego, CA H.R. D'Agostino, MD • A.J. Carrillo, MD S. Kim, MD • Y Lee, MD • R.M. O'Laoide, MB, et al PURPOSE: We evaluated the effect of percutaneous intratumoral alcohol injection in rabbits with VX2 carcinoma in the liver and kidney. MATERIALS AND METHODS: Twenty New Zealand White rabbits underwent US-guided percutaneous implantation ofVX2 carcinoma in the liver and kidney. Tumors were treated with US-guided intratumoral alcohol injection (n = 12). The control group received intratumoral injection of 0.5 mL of normal saline (n = 8). The effect of intratumoral alcohol injection was evaluated by measuring the duplication time and length of survival, at postmortem examination and at histopathologic study of the specimens. RESULTS: Control animals had a duplication time of 2.65 days and lived 9 days. Treated animals had a duplication time of 5.3 days and lived 15 days. Lung metastasis caused major distress. Histopathologic studies of treated tumors showed tumor necrosis and infarction of peritumoral liver. CONCLUSION: Intratumoral alcohol injection caused a direct lethal effect on the tumor and peritumoral liver, resulting in delayed tumor duplication time and prolonged survival in treated animals.


1:45 PM

CT-guided Percutaneous Tumor Ablation with Alcohol J.F. Botet, MD, New York, NY • L. Blumgart, MD N. Kemeny, MD • J. Fortner, MD PURPOSE: We evaluated the response to percutaneous intralesional injection of alcohol into liver tumors. MATERIALS AND METHODS: Eighteen patients with unresectable disease have undergone treatment during the past 4 years. Of 10 patients with hepatoma, seven had undergone previous lobar resection and presented with recurrences; treatment was primary in three patients with Child stage C disease. Eight patients with metastatic colorectal cancer to the liver, which had been treated previously with lobar resection, were also treated. Biopsy of all lesions was performed prior to treatment. We used aCT-based stereotactic method that specifically addresses the margins of the lesion. The total volume is calculated by using the average maximum diameter; this is given over three sessions on alternating days. A 22-gauge metal needle is used. RESULTS: Eight of 10 patients with hepatocellular cancer are long-term responders, with follow-up of 6-24 months (average, 12 months). This has been documented with follow-up CT, a-fetoprotein levels, and, in five cases, repeated biopsies of the lesions. In one patient who has undergone three treatment courses over an 8-month period, a-fetoprotein levels did not decrease although repeat biopsies and CT and gallium scans were negative. One other patient developed a second recurrence and elected to undergo liver transplantation. The average number of lesions was 2.5. The largest measured 6 em. All eight patients with metastatic colorectal disease have responded, with an average follow-up of 7 months. This was documented with CT, carcinoembyonic antigen level, and, in two cases, repeat biopsies. The average number of lesions was two. The largest was 5 em. CONCLUSION: The technique appears to be a promising tool in the treatment of recurrent or metastatic disease and will be discussed in detail. 153


High Sensitivity of Selective Venous Sampling Catheterization for Localization of Persisting Medullary Thyroid Carcinoma after Initial Surgery T. de Baere, MD, Villejuif, France N. Abdelmoumene, MD • A. Roche M. Schlumberger, MD PURPOSE: We evaluated selective venous sampling (SVS) for detection and localization of residual disease or distant metastases in patients with persistent high

56 • Journal of Vascular and Interventional Radiology January February 1994

calcitonin levels after adequate surgery for medullary thyroid carcinoma. MATERIALS AND METHODS: Nineteen patients underwent SVS for serum calcitonin measurement. Follow-up after SVS was 5.3 (0.2-12.9) years. RESULTS: One patient with a normal follow-up (1.6 years) had no calcitonin gradient between selective samples and peripheral level. Gradients, ranging from 1.3 to 8.2, were found in 18 patients. All had gradients localized to the neck or upper mediastinum, and five patients had additional distant gradients. Neck or upper mediastinum gradients were confirmed at surgery (n = 12) or CT (n = 6). The patient with the lowest gradient value did not evidence tumor at surgery. Imaging work-up at the time of sampling did not evidence tumor at the site of the five distant gradients. However, at the site of those gradients, distant metastases appeared on CT during the follow-up (medial delay, 2.9 years). CONCLUSION: SVS seems to be a sensitive method for the detection of persisting medullary carcinoma. Furthermore it allows detection and localization of distant spread before any imaging technique. 154 • 2:15 PM Intraarterial Neoadjuvant Chemotherapy of Soft-Tissue Sarcomas

D.M. Coldwell, MD, Seattle, WA • C. Collins, MD E. U. Conrad, MD • S.J. Althaus, MD N.H. Patel, MD • D.J. Glickerman, MD PURPOSE: We determined the effectiveness of intraarterial chemotherapy with cisplatin (60 mg/m 2 on days 1 and 8) combined with adriamycin (60-90 mg/ m 2 ) given intravenously over 96 hours for three cycles prior to surgical resection. MATERIALS AND METHODS: Twenty-five patients with primary soft-tissue sarcomas received the above treatment regimen. Tumor pathology before chemotherapy and at surgery, amount of tissue necrosis, angiographic and cross-sectional imaging changes, and clinical response were noted. RESULTS: Of the 25 patients, 17 (68%) had improvement in pain or function. Tumor volume decreased in 13 (52%) but did not correlate with hypovascular changes seen at angiography. Median viable tumor present after this therapy was 43%, but 11 (44%) had less than 15% viable tumor. The projected disease-free and overall survival at 24 months are 80% and 88%, respectively. CONCLUSION: This intraarterial therapy is an effective means to control soft-tissue sarcomas prior to surgical resection and compares well with other regimens.

155 • 2:30 PM Clinical Trial of the Bard CT Guide System

K.T. Brown, New York, NY • J.F. Botet, MD G.I. Getrajdman PURPOSE: A clinical trial is in progress to evaluate the utility, safety, and efficacy of the recently introduced Bard CT guide system. MATERIALS AND METHODS: One hundred consecutive patients were entered into the study to undergo biopsy with use of an 18-gauge Bard Monopty device with the CT guide system followed by biopsy with conventional 18-gauge coring needle. Thirtyeight percent of biopsies were of the liver, 22% were retroperitoneal, 20% were renal/adrenal, 9% were pelvic, 8% were pancreatic, and 3% were miscellaneous. Chest biopsies were excluded. The group of patients able to undergo biopsy with both devices was compared with a historical control group of 45 patients who underwent biopsies with the conventional method only and the subgroup of patients entered in our trial in whom use of the Bard device was technically not feasible. We recorded the duration of the procedure, accuracy of needle placement, results of the biopsy, and complications. RESULTS: To date, there is no significant difference in the time required to perform a biopsy with the CT guide system compared with conventional techniques. The positioner is very accurate, delivering the needle tip to exactly the preprogrammed location within ± 2° of the programmed angle in 95% of cases. The results of the two techniques are concurrent in 84% of cases. In 14% of cases the specimen from the Bard needle was negative, whereas that obtained with the coring needle was positive for malignancy. In 2% of cases the Bard needle specimen was positive for malignancy, and the coring needle specimen was negative. No complications were associated with use of the device. CONCLUSION: There is a relatively steep learning curve associated with use of the device; however, once the physicians and technologists gain familiarity with the device the time required to perform a biopsy is almost identical to that with conventional techniques. There are several technical limitations to the use of this device related to patient size, gantry size, location of the lesion, amount of tissue obtained, and angle of approach necessary for the biopsy. Overall, the device could not be used on 30% of the patients in our study, most commonly related to the need for a lateral approach. The incidence of false-negative results with the Bard device is likely secondary to sampling error, as larger tissue volumes were obtained with the 18gauge coring needle.

1994 SCVIR Meeting Abstracts • 57 Volume 5 Number 1

156 • 2:45 PM Optical Biopsy with Use of a Raman Spectroscopic Fiberoptic Needle for in Situ Malignancy Detection D. C. Redd, MD, Philadelphia, PA • C.J. Frank, MS R.L. McCreery, PhD • T.S. Gansler, MD PURPOSE: Needle biopsy is subject to misrepresentative sampling and thus inaccurate diagnosis. Tech~i­ cal advances in Raman spectroscopy now make rapId characterization of tissues possible. The spectra of human breast tissues were recorded with a Raman spectroscopic fiberoptic needle to evaluate the efficacy of an optical biopsy system. MATERIALS AND METHODS: A single-stage spectrograph with CCD detector and near-infrared excitation was used for recording Raman spectra (n = 1 000) from normal and neoplastic breast tissues. A fiberoptic probe was utilized for in situ detection in a tissue model of breast cancer. RESULTS: The Raman spectrum of normal breast tissue is highly reproducible with principal peaks near 1,072,1,285,1,439, and 1,654, with 1,745 C~-l reflecting the presence of lipids. In breast carcmo~a, the lipid features are lost, with new features attributable to collagen (type I and III). In a breast cancer model the Raman spectrum changed from normal Oipid)'to neoplastic (collagen) as the fiberoptic needle passed into the area of tumor. CONCLUSION: Raman spectroscopy is useful for characterizing the molecular components of tissues. By recording sequential spectra in real-time as it passes through tissue planes, the fiberoptic needle may be useful for in situ optical biopsy. Additional studies are ongoing to correlate Raman spectral and immunohistochemical data and to further refine the design of a fiberoptic needle for optical biopsy.

Thursday Afternoon • 6D • Scientific Session 27 Embolotherapy MODERATOR:

Arthur C. Waltman, MD, Boston, MA

157 • 1:30 PM Endovascular Treatment of Giant Intracranial Aneurysms R. T. Higashida, MD*, San Francisco, CA V. V. Halbach, MD* • C.F. Dowd, MD G.P. Teitelbaum, MD • C.G. McDougall, MD G.B. Hieshima, MD* PURPOSE: We assessed the feasibility, technical strategies, results, and long-term outcome of patients with giant intracranial arterial aneurysms treated by means of interventional neurovascular therapy. MATERIALS AND METHODS: Of 321 total cases in which treatment was undertaken, 95 cases (29.6%) were considered giant intracranial aneurysms ( > 2.5 cm). In 74 patients the aneurysm was located in the anterior circulation and in 21 cases, in the posterior circulation. The clinical presentation of these patients was mass effect in 81 cases (85.3%), subarachnoid hemorrhage in 11 cases (11.6%), and thromboemboli in three cases (3.2%). For occlusion of the parent vessel, test occlusion, followed by occlusion proximal to or across the aneurysm neck, was performed with use of detachable balloons. For direct aneurysm occlusion, detachable balloons, fibered coils, or electrolytic detachable coils were used. RESULTS: In 64 cases (67.4%), the aneurysm was treated with parent artery occlusion. In 31 cases (32.6%) direct treatment of the aneurysm with preservation of the parent artery was possible. Complications related to treatment included stroke in 11 cases (11.6%), transient cerebral ischemia in six cases (6.3%), and death in nine cases (9.5%). CONCLUSION: Treatment of giant intracranial aneurysms in the anterior and posterior cerebral circulations is technically feasible with use of interventional techniques. 158 • 1:45 PM Sclerotherapy of Vascular Malformations and Hemangiomas in Children R.D. Kaye, MD, Pittsburgh, PA • R. Towbin, MD C. Fritz, MD PURPOSE: Embolotherapy and sclerotherapy, alone or in combination, have become increasingly important as primary or adjunctive therapy in children with vascular malformations and hemangiomas. This re-

58 • Journal of Vascular and Interventional Radiology January-February 1994

port discusses the indications, effectiveness, and safety of sclerotherapy and proposes an approach for children with these lesions. MATERIALS AND METHODS: Over the past 12 years 197 vascular interventions have been performed in 152 children ranging in age from 1 day to 20 years. A subgroup of 15 children have been treated with sclerotherapy. Twenty-five procedures have been performed with use of absolute alcohol in nine girls and six boys ranging in age from 2 to 17 years. Areas involved included face (n = 4), lip (n =' 3), neck (n = 2), cheek (n = 2), and extremity (n = 2) and lymphangiomas of the chest wall, mediastinum, and hip. All procedures were image-guided and required one to three injections of ethanol. MR and/ or direct puncture angiography was performed on all patients. RESULTS: Ethanol injection resulted in an improvement in size and color in six patients and cure in seven patients. One patient was lost to follow-up, and results are too early to evaluate in one patient. Significant swelling was present in the majority of cases. Aside from a single instance of transient facial weakness, no complications have been seen. CONCLUSION: Sclerotherapy is effective and safe as both a primary and adjunctive therapeutic measure for treatment of slow-flow capillary or venous malformations. Our preliminary results with treatment of lymphangiomas are exciting and this may become a useful therapeutic strategy, but further experience is necessary. 159 • 2:00 PM Ethanol Endovascular Ablation of Brain Arteriovenous Malformations: Initial Results W.F. Yakes, MD, Englewood, CO • L.E. Krauth, MD J.N. Dreisbach, MD • C.N. Seibert, MD PURPOSE: We determined the safety and efficacy of undiluted absolute ethanol as a primary mode of therapy in the management of intraaxial brain arteriovenous malformations (AVMs). MATERIALS AND METHODS: Six patients (three male, three female) underwent transcatheter ethanol devascularization of their brain AVMs (age range, 19-54 years; mean, 38 years). Clinical presentations included subarachnoid hemorrhage, headache, seizure, and vertigo. Superselective catheterizations were performed. Amytal testing preceded all ethanol procedures (eight procedures). Follow-up consisted of clinic visits, MR imaging, and arteriography. Two patients were lost to follow-up after 3 months (follow-up range, 3-13 months; mean, 6 months). Neuroleptic anesthesia was used in five patients; general anesthesia was used in one. RESULTS: No recanalizations or neovascular recruitment was observed at follow-up in any patient. When comparing the immediate postembolization arterio-

grams to follow-up arteriograms, progressive AVM thrombosis routinely occurs. One patient with an occipitallobe hemorrhage and resultant homonymous hemianopsia spontaneously recovered his vision on the angiographic table. A 50% complication rate was observed (one patient in whom an episode of dizziness totally resolved; one patient with a reading difficulty totally resolved; one patient with bilateral upper outer quadrant visual field deficits, improving). CONCLUSION: Our initial results indicate that ethanol can be used safely in the brain, and progressive AVM occlusion is a common feature at follow-up. 160 • 2:15 PM Long-term Results of Bronchial Artery Embolization for Hemoptysis in Benign Disease C. Teeuwen, MD, Blaricum, The Netherlands A Algra, MD • T. T. Overtoom, MD PURPOSE: We determined the long-term outcome of bronchial artery embolization (BAE) in patients with hemoptysis. MATERIALS AND METHODS: In 155 patients with hemoptysis, bronchial arteriography was performed. BAE was performed in 47 patients, 85 patients received conservative treatment, and 23 patients underwent surgery. Long-term follow-up of more than 24 months was completed in 140 patients (90.4%). Mean follow-up was 49.7 month. RESULTS: Mter 2 years, 70% of patients in the BAE group, 78% in the surgical group, and 55% of patients treated conservatively were free of hemoptysis. After 5 years results were 59% for BAE (95% CI, 39%-74%), 63% for surgery (34%-82%), and 46% for conservative treatment (34%-58%). The risk of rebleeding was 32% (95% CI, 18%-61%) lower in the BAE group than in the group treated conservatively. This percentage hardly changed after adjustment for sex, age, and signs of hypertrophy at angiography. CONCLUSION: These data indicate that BAE can be beneficial in the treatment of hemoptysis; however, no statistical significance was reached. 161 • 2:30 PM Hemoptysis in Patients with Pulmonary Tuberculosis: Treatment with Transcatheter Embolization R.A Alcantara-Peraza, Mexico DF, Mexico AG. del Campo, MD • Y. Kimura, MD H. Ferral, MD • M. Cardoso, MD J.L. Hernandez, MD PURPOSE: We describe the angiographic findings in patients with pulmonary tuberculosis and report our experience in the treatment of hemoptysis with transcatheter embolization.

1994 SCVIR Meeting Abstracts • 59 Volume 5 Number 1

MATERIALS AND METHODS: Thirty-five patients with proved pulmonary tuberculosis and massive hemoptysis (500 mL over 24 h) were treated. Bronchoscopy was performed before angiography to localize the bleeding site. Bronchial arteries were selectively catheterized with 5- or 6-F Sidewinder catheters. Embolization was performed if enlarged bronchial arteries, fistulas, or active bleeding sites were found. Gelfoam plugs were used until complete cessation of flow was demonstrated. All patients received a full course of antituberculous drugs starting before or immediately after embolization. Patients have been followed up clinically for a mean period of 3 years. RESULTS: Catheterization of the bronchial arteries was successful in all patients. Twenty-seven had separate origins for right and left arteries, and eight patients had a common trunk. Immediate control of hemoptysis was achieved in all patients. Rebleeding was seen in five patients (14%) and embolization was repeated. Thirty patients (85%) presented with fever, chest pain, and dysphagia after embolization; symptoms disappeared in 48-72 hours. CONCLUSION: Transcatheter embolization is a safe and effective method to control massive hemoptysis in patients with pulmonary tuberculosis. Treatment must be completed with a full course of antituberculous drugs. 162 • 2:45 PM Evaluation of a Detachable Silicone Balloon for Noncerebral Arterial Applications

S.O. Trerotola, MD, Indianapolis, IN R.N. Messersmith, MD • M.E. Hicks, MD A. van Breda, MD • A.M. Vinson, MD A.B. Zajko, MD PURPOSE: Detachable balloons may offer advantages over other embolic agents for arterial embolization. Since the withdrawal of the Bard-Parker Mini-balloon from the market, there have been no FDA-approved detachable balloons available in the United States. The purpose of this study was to evaluate a detachable silicone balloon for noncerebral arterial applications. MATERIALS AND METHODS: Fourteen arterial embolizations were performed with use of detachable balloons alone (n = 5) or with coils (n = 9) in 13 patients. Indications for embolization included gastrointestinal bleeding (n = 2), hemobilia (n = 1), arteriovenous fistula (n = 6), aneurysm (n = 2), and splenectomy (n = 2). Indications for balloon use included high flow, need for rapid occlusion, and operator preference. RESULTS: Balloon embolization was successful in 10 patients and failed in three. In two of the latter, coil embolization was successful. Three major complications-premature balloon detachment with migration to the lung-occurred. Two balloon ruptures and two

late balloon deflations occurred without clinical significance. All but one embolization procedure resulted in clinical success, with continued success at 6-month follow-up. CONCLUSION: The detachable silicone balloon is an effective arterial embolic agent. Balloon rupture and/ or deflation is possible when adjacent coils are used, but does not adversely affect the outcome of the procedure. Premature detachment may occur in highflow situations.

Thursday Mternoon • 6E • Scientific Session 28 Biopsy and Drainage MODERATOR:

Thomas M. Vesely, MD, St. Louis, MO

163 • 1:30 PM Predictive Value of the Gram Stain

B.F. Stainken, MD, San Diego, CA • D. Bickmore H.R. D'Agostino, MD PURPOSE: We attempted to determine the clinical utility of the Gram stain as a predictor of culture results. MATERIALS AND METHODS: Each element of 250 Gram stain readings on 184 interventional radiology patients was codified and compared with the site of aspiration, the described nature of the aspirate, its volume, and the culture results. RESULTS: There is no significant (P < .05) correlation identified between findings of bacteria or white blood cells on Gram stain and culture results. Specifically, findings for large quantities of polymorphonuclear cells or findings of bacteria were not predictive of culture results. The absence of bacteria and white blood cells on Gram stain, however, in 11.2% of stains (28 of 250) was never associated with culture positivity (P < .02). CONCLUSION: The routine use of an immediate Gram stain in determining the need for percutaneous drainage versus aspiration is based on an assumption that there is an association between Gram stain and culture results. The only significant association, however, is that for the absence of white cells or bacteria on staining and no growth on cultures (P < .02).

60 • Journal of Vascular and Interventional Radiology January-February 1994

164 • 1:45 PM Transjugular Liver Biopsy: Comparison of a New Cutting Needle and the Aspiration Needle T. Sasson, MD • P.C. Lakin, MD, Portland, OR R.E. Barton, MD • H.A. Timmermans, BS B. Uchida, BS • F.S. Keller, MD PURPOSE: A new cutting needle is compared with the aspiration needle for transjugular liver (TJL) biopsy. MATERIALS AND METHODS: We retrospectively evaluated 40 patients who underwent a TJL biopsy performed at our institution. Twenty patients underwent an aspiration TJL biopsy performed with the Ross transseptal needle, and 20 underwent TJL biopsy performed with a spring-loaded, side-port cutting needle developed in the Dotter Institute research laboratory. Both needles are 18 gauge and are guided through a 9-F catheter. Comparisons were made for sex, age, indications for biopsy, success rate, quality of biopsy specimen, and complication rate. RESULTS: Although the groups were similar for patient demographics, the success rate was greater with the cutting needle (95% vs 80%) than with the aspiration needle. All specimens were diagnostic, but the cutting needle specimens were less fragmented (7% vs 25%) and of greater length (ie, longer than 8 mm) (63% vs 25%) than the aspiration needle specimens. All interventionalists who had experience with both needles considered the cutting needle easier to use. No significant complications occurred. CONCLUSION: The cutting TJL biopsy needle offers improved ease of use, biopsy success rate, and quality of biopsy specimen compared with the aspiration TJL biopsy needle. 165 • 2:00 PM CT·guided Removal of Osteoid Osteoma in Children: A Different Approach R.B. Towbin, MD, Pittsburgh, PA M. Moreland, MD • R. Kaye, MD • C. Fritz, MD PURPOSE: Osteoid osteoma is a benign bone tumor found in adolescents and young adults and was first described by Jaffe in 1935. Traditionally, a wide surgical resection is recommended to reduce the potential of local recurrence. Incomplete incision is often the result of poor intraoperative visualization of the nidus. CT-guided removal of osteoid osteoma has potential advantages over operative removal. We have devised a simple guided approach for removal of the tumor nidus. MATERIALS AND METHODS: Four children ranging in age from 7 to 14.5 years (x = 11 years) pr~­ sented with typical signs and symptoms of osteOId osteoma. Two lesions involved the proximal femur and two, the proximal tibia. A Steinman pin is inserted

into the bone for guidance. A 3- or 5-mm trephine is coaxially positioned over the pin into the lesion, and the nidus cored out. RESULTS: Five tracts were needed to remove the four niduses. Postoperatively, all children were immediately pain free. No postoperative complications occurred. The children have been followed up for a mean of 8 months (2-12 months) with no evidence of recurrence. CONCLUSION: CT-guided removal of osteoid osteoma appears to be a safe and effective method. There are advantages over surgical removal including a lower potential for postoperative stress fracture, avoidance of the need for a bone graft, shorter convalescent period, and confirmation of nidus removal. 166 • 2:15 PM Percutaneous Drainage of Suppurative Adenitis in Infants and Children S.A. Connolly, MD, Boston, MA • F.A. Hoffer, MD PURPOSE: The purpose of this study was to make interventional radiologists aware that suppurative cervical adenitis (SCA) can be treated with percutaneous US-guided drainage as an alternative to operative incision and drainage. MATERIALS AND METHODS: In a retrospective study, results for seven children with SCA age 2 1/ 2 months to 18 months who underwent percutaneous drainage from 1991 to 1992 were compared with those for 21 children who underwent surgical incision and drainage of SCA from 1988 to 1992. Percutaneous drainage was preceded by acquisition of an US scan demonstrating suppuration. With the patient under general anesthesia, an 8.5-F pigtail catheter (Cook) with a l-cm-diameter locking pigtail, was introduced with use of US and fluoroscopic guidance and Seldinger technique. RESULTS: With percutaneous technique, an average of 2.5 mL of pus was aspirated initially and an average of 9 mL of material drained within 48 hours. The average time of catheter drainage was 3 days. Cervical node size decreased markedly during drainage and returned to normal within 1-2 weeks. The length of hospital stay after intervention was the same with the two approaches. CONCLUSION: Advantages offered by percutaneous drainage include a smaller and more cosmetic incision, continuous drainage, and the ability to perform the procedure in a radiologic suite instead of a costly operatingroom.

1994 SCVIR Meeting Abstracts • 61 Volume 5 Number 1

167 •


Fine-Needle Aspiration of Pancreatic Masses: Diagnosis of Pancreatic Carcinoma with Detection of the K-ras Mutation F.P. Boudghene, MD, Paris, France • T. Urban, MD J.F. Bemaudin, MD • F. Breittmayer, MD S. Ricci, MD • J.M. Bigot, MD PURPOSE: Detection of a codon 12 mutated K-ras gene in samples obtained at fine-needle aspiration (FNA) of pancreatic lesions was prospectively evaluated for the diagnosis of pancreatic carcinoma. MATERIALS AND METHODS: During a 2-year period, 20 consecutive patients (12 male, eight female; mean age, 66 years) with a pancreatic masslike lesion underwent FNA (22-gauge needle) with US or CT guidance. Mutation was detected by means of polymerase chain reaction and restriction fragment length polymorphism (PCR-RFLP) analysis, and results were obtained within 48 hours. RESULTS: DNA evaluation was feasible in 16 of the 20 samples, and a mutated K-ras gene was detected in 11 of 12 definitive pancreatic carcinomas (92%). No mutation was observed in chronic pancreatitis (n = 3) or islet cell tumor (n = 1). The correct diagnosis was obtained by means of pathologic examination alone of the samples obtained at FNA in 13 of 20 samples (65%). When combined with the detection of a mutated K-ras, the diagnosis yield was 16 of 20 (80%) and 16 of 16 (100%) when adequate material was sampled. CONCLUSION: Screening of pancreatic samples obtained at FNA for K-ras mutation with use of PCRRFLP analysis could constitute a particularly helpful tool for the diagnosis of pancreatic masses. 168

2:45 PM

Thyroid Biopsies: Fine-Needle Aspiration Biopsy versus Spring-activated Core Biopsy Needle in 102 Patients S.F. Quinn, MD, Portland, OR • H.A Nelson, MD T.A Demlow, MD PURPOSE: This report represents our experience with sonographic guidance for thyroid biopsies with a single-action spring-activated core biopsy needle, and we compare this technique with fine-needle aspiration biopsy (FNAB). MATERIALS AND METHODS: This was a prospective evaluation of 102 patients who underwent sonographically guided thyroid biopsies. The biopsy results from the 21-gauge FNAB (n = 102) were compared with results from 18-gauge (n = 1), 19-9auge (n = 2), 20-gauge (n = 43), or 21-gauge (n = 56) single-action spring-activated core biopsy needles (Temno; Bauer Medical, Tampa, Fla).

RESULTS: If the suspicious and diagnostic results were combined, the diagnostic accuracies would be the following: Temno 90.2% (92 of 102), FNAB 80.3% (82 of 102), and the combination of both needles 97.1% (99 of 102). The complication rate was 0.98% (one of 102). This complication was a minor soft-tissue infection successfully treated with oral antibiotics. CONCLUSION: In summary, sonographically guided biopsies of the thyroid performed with single-action core biopsy needles are safe and effective. The results with these needles are better than the results with FNAB, but the best results come from use of both needles in the same patient.

Thursday Mternoon • 6C • Scientific Session 29 Tracheal and Enteric Stents Alan H. Matsumoto, MD, Charlottesville, VA


169 •


Benign Bronchial Stenoses: Treatment with Palmaz Stents AH. Matsumoto, MD, Charlottesville, VA C.J. Tegtmeyer, MD • C.E. Rose, MD T. Daniel, MD • R. Fingland, DVM C.G. Tribble, MD, et al PURPOSE: We evaluated the efficacy of the Palmaz stent for the treatment of benign bronchial stenoses. MATERIALS AND METHODS: Eight patients underwent placement of 14 Palmaz stents for treatment of nine bronchial stenoses with use of fluoroscopic guidance. Eight lesions were at transplant bronchial anastomoses and one stenosis was secondary to histoplasmosis. Indications for stent placement included increasing dyspnea following lung transplantation in three patients and failure to wean from a ventilator in five patients. All patients had bronchoscopic and bronchographic evidence of bronchial stenosis. Mter stent placement, patients were followed up with serial bronchoscopy and pulmonary function tests. RESULTS: Following stent placement, all three patients treated for dyspnea and three of five patients treated for ventilator dependence improved significantly. Ingrowth of granulation tissue occurred in three patients at 6-15 weeks, requiring transbronchoscopic electrocautery and repeated balloon dilation of the stents. Including the three retreatments, six patients (75%) are doing well at 4-16 months offollow-up. The two clinical failures occurred in lung transplant recipients who died of overwhelming pneu-

62 • Journal of Vascular and Interventional Radiology January-February 1994

monia. In one of these patients, two stents were traumatized and deformed by overzealous suctioning. In the other clinical failure, two stents were dislodged during endoscopic bronchial lavage. CONCLUSION: The Palmaz stent appears to be a useful device for treating benign bronchial stenoses. Although ingrowth of granulation tissue through the stent does occur, it can be managed with transbronchial electrocautery and repeated stent dilation. 170 • 4:00 PM Tracheobronchial Stent Placement: Three-year Experience M. Maeda, MD, Nara, Japan • H. Uchida, MD H. Otsuji, MD • S. Iioka, MD • S. Makutani, MD K. Kichikawa, MD, et al PURPOSE: The initial and midterm efficacy of the Z stent in the tracheobronchus was evaluated. MATERIALS AND METHODS: Thirty-two patients with tracheobronchial stenosis were treated with placement of modified Z stents (13 cases) and spiral Z stents (19 cases). The stenoses were due to eight benign lesions (posttracheotomy, postsurgical stenosis, tuberculosis, and bronchomalacia) and 23 malignant lesions. Follow-up studies included use of the HughJones scale, plain radiography, CT, bronchoscopy, and radionuclide examinations. RESULTS: Dyspnea improved in 28 of 32 patients immediately after stent placement. Eight patients (five with benign and three with malignant lesions) could be followed up over 1 year up to 3 years (two died at 1 year 3 months and 1 year 8 months, respectively; six are alive). Four restenoses due to tumor ingrowth and granulation were seen but could be treated with repeated stent placement, laser use, or injection of a steroid. Two breakages of stents occurred without any symptoms. No serious complications were encountered during follow-up periods. CONCLUSION: Z-stent placement in tracheobronchial stenoses was effective to relieve dyspnea. Prognoses in patients with malignancy depend on tumor control; however, midterm follow-up results show the safety and effectiveness of Z stents, and their clinical use is encouraged. 171 • 4:15 PM Expandable Stents for Treatment of Tracheobronchial Lesions B.D. Petersen, MD, Portland, OR • R.E. Barton, MD B. Uchida, BS* • R.R. Saxon, MD • J. Rosch, MD* F.S. Keller, MD PURPOSE: Expandable metallic stents were evaluated for treatment of malignant and benign tracheobronchial stenoses.

MATERIALS AND METHODS: Seven patients, aged 44- 73 years, were treated with placement of expandable metallic stents for severe dyspnea. Three patients with malignant lesions at the level of the carina received Gianturco-Rosch Z (GRZ) stents (two silicone covered, one noncovered). Three patients with benign tracheal lesions (one with tracheomalacia, two postoperative) received uncovered GRZ stents. One patient with a postoperative bronchial stricture received a noncovered Wallstent. RESULTS: All patients with malignant lesions improved symptomatically after stent placement and remained without significant dyspnea until their death (2 weeks and 5 months) or until present (3 months). One patient with benign stricture and one with tracheomalacia also responded well to stent placement. One has been asymptomatic for more than 12 months; another patient with severe chronic obstructive pulmonary disease improved significantly but died of endstage disease 5 months later. In the other two patients with benign lesions, stent placement did not have satisfactory results. A GRZ stent in one patient was not placed properly in a high tracheal lesion, and a Wallstent placed in a bronchial lesion resulted in obstructive pneumonia. Both stents were removed surgically. CONCLUSION: Expandable metallic stents offer effective palliation of malignant tracheal bronchial stenoses and may be effective in treatment of benign lesions. 172 • 4:30 PM Clinical Efficacy of Covered Z Stents in the Treatment of Malignant Esophageal Obstructions and Esophagorespiratory Fistulas R.R. Saxon, MD, Portland, OR • R.E. Barton, MD P.C. Lakin, MD • W.C. Wu, MD • R.M. Katon, MD J. Rosch, MD* PURPOSE: The efficacy of Z stents was evaluated in the treatment of malignant esophageal obstructions and esophagorespiratory fistulas. MATERIALS AND METHODS: Self-expanding, metallic, covered Gianturco-Rosch Z (GRZ) stents were placed in 48 patients with high-grade malignant esophageal obstructions (n = 39) and esophagorespiratory fistulas (n = 9). RESULTS: Long-term relief of dysphagia was achieved in 34 of 39 patients with esophageal obstructions (87%). Two patients failed to respond, and three patients developed recurrent dysphagia during followup. Six of the nine patients with esophagorespiratory fistulas were able to resume a normal diet, while two of three patients not able to resume a normal diet could eat solids without aspirating. Survival after stent placement ranged from 1 week to 2 years (mean, 3.5 months). The majority of patients died secondary to a progression of their malignancy. Complications

1994 SCVIR Meeting Abstracts • 63 Volume 5 Number 1

occurred during follow-up in 12 patients and included stent migration (five patients), food impaction (two patients), tumor ingrowth (one patient), hematemesis (one patient), esophageal perforation with sepsis (one patient), and pressure necrosis with hemorrhage (two patients). The latter three patients died. Eight of the 12 complications were minor and without sequelae. CONCLUSION: Covered GRZ stents are an effective means of palliating patients with malignant esophageal obstructions and esophagorespiratory fistulas. Clinical efficacy compares favorably with results of published series of rigid endoprosthesis placement. 173

4:45 PM

Silicone-covered Stents for Upper Gastrointestinal Tract Obstructions with and without Fistula M. Maeda, MD, Nara, Japan • H. Uchida, MD R. Nishigori, MD • H. Yoshimura, MD H. Otsuji, MD • K. Kichikawa, MD PURPOSE: The authors attempted to establish the efficacy of silicone covered stents for upper gastrointestinal tract obstructions with or without fistula. MATERIALS AND METHODS: Twenty-two patients with upper gastrointestinal tract obstructions due to esophageal cancer (11 cases), esophago-cardial cancer (two cases), recurrent cancer at anastomoses (five cases), and postoperative scarring at anastomoses (four cases), were treated with silicone covered stents. Two type of stents, modified Z stents with hooks (seven cases) and spiral Z stents with flared ends that were 4-5 mm larger than the middle (15 cases), were used. RESULTS: Stents were successfully placed in 22 cases. Stents dilated the obstruction and maintained patency during follow-up periods up to 11 months. Stent alone failed to close fistulas completely. Glue was injected outside the stent to close the fistula in two cases. Complications of mediastinitis (n = 1), esophagotracheal fistula formation (n = 1), and reflux esophagitis (n = 1) were encountered. Stents without hooks migrated in four cases. Stents with the flared ends placed in four cases have not migrated. CONCLUSION: Silicone covered stents offer a safe and effective treatment for upper gastrointestinal tract obstructions. Stent placement and glue injection were effective in closing gastrointestinal tract fistulas.



Evaluation of Nitinol Stents for Treatment of Malignant Esophageal Obstructions H.J. Wagner, MD, Marburg, Germany D. Bartsch, MD • K.J. Klose, MD PURPOSE: The authors evaluated newly developed self-expanding nitinol stents in the treatment of malignant dysphagia. MATERIALS AND METHODS: From June 1992 through August 1993, 14 consecutive patients (mean age, 68.4 years ± 11.3) with malignant dysphagia due to inoperable obstructions of the esophagus were prospectively treated with insertion of 15 self-expanding nitinol stents (Ultraflex). RESULTS: Stent placement was successful in all 14 patients. One stent expanded insufficiently and was inadvertently removed during retrieval of the application system. Mean time until complete opening of the stent was 5.9 days ± 3.1. In five of 14 patients difficulties of stent deployment were observed. Mean dysphagia score (0-4) decreased from 2.8 ± 0.7 to 0.5 ± 0.7 (P < .005) after stent insertion; activity score (0-4) decreased from 2.5 ± 1.1 to 1.3 ± 0.5 (P < .005). Complications related to stent insertion were not noted. During the mean observation time of 132 days ± 103, three patients experienced recurrent dysphagia: two due to bolus occlusion, one due to tumor ingrowth through the mesh. Mean survival time was 190 days ± 97. CONCLUSION: Disadvantages ofnitinol stents are their low expansion forces. Modifications of the design as well as a complete coating are still required to improve short- and long-term results.

Thursday Mternoon • 6F • Scientific Session 30 Chemoembolization Michael J. Pentecost, MD, Philadelphia, PA



3:45 PM

Chemoembolization of Hepatocellular Carcinoma with Doxorubicin and Ethiodol K.R. Stokes, MD, Boston, MA • K. Stuart, MD R. Jenkins, MD • M.E. Clouse, MD • C. Trey, MD PURPOSE: The efficacy and safety of hepatocellular chemoembolization were evaluated. Chemoembolization involved superselective arterial injection of doxorubicin/Ethiodol followed by Gelfoam powder embolization.

64 • Journal of Vascular and Interventional Radiology January-February 1994

MATERIALS AND METHODS: Eighty patients including 12 patients with portal vein occlusion, u~der­ ,,:ent 88 chemoembolization procedures. SeventyeIght percent of patients had cirrhosis. RESULTS: Thirty-day mortality was 17%. Eighty percent of patients had more than 25% decrease in tumor mass. Overall, 1- and 2-year survival rates were 63% and 42%, respectively, with median survival of 16 months, which compares extremely well with historic median of6-14 weeks in similar patients. CONCLUSION: This form of chemoembolization has been shown to be an effective treatment of hepatocellular carcinoma in a non-Asian population with improvement in survival and quality of life compared with conventional therapy. In patients with portal vein occlusion, 30-day mortality was 17%, with 35% 12-month survival. In those patients with normal bilirubin levels, 30-day mortality was 4%, with median survival of 21 months, 12-month survival of 8%, and 24-month survival of 54%. 176 •


Results of Hepatic Arterial Embolization for Metastatic Tumors D.M. Coldwell, MD, Seattle, WA • S.J. Althaus, MD D.J. Glickerman, MD • N.H. Patel, MD PURPOSE: We evaluated the efficacy of hepatic artery embolization (RAE) without chemotherapy for the treatment of metastatic tumors. MATERIALS AND METHODS: Fifty patients with metastatic tumors to the liver, 23 of 50 (46%) from colorectal carcinoma, six of 50 (12%) from carcinoid were treated with 5-fiuorouracil/leucovorin for 6 ' weeks before RAE with polyvinyl alcohol (150-250 j..lm).

RESULTS: Forty-three of the 50 patients (86%) had symptomatic improvement. The metastases stabilized in 24 patients (48%). The mean survival time was 16 months, with a mean time to recurrence of 24 weeks. Two complete responses and four partial responses were present. Complications were present in four patients; three had gallbladder infarction, and one had arterial dissection. CONCLUSION: The majority of patients have subjective improvement after RAE. Although radiologic responses are not common, most patients have stabilization of disease seen at CT. The procedure is relatively safe. The results of RAE without the addition of chemotherapy should be used as a reference with which to compare hepatic arterial chemoembolization.

177 •

4:15 PM

Comparative Study of Different Interventional Therapies for Primary Hepatic Carcinoma: Analysis of 468 Cases Y.C. Jia, MD, Shanghai, China • Q. Liu • Z. Wang J. Tian • H. Ye • F. Wang PURPOSE: We compared three types of treatment in the management of late-stage hepatocellular carcinoma (RCC). PATIENTS AND METHODS: A total of 468 patients with late-stage RCC were randomly allocated to three different therapy regimens: chemotherapy alone with mitomycin C, Adriamycin, and 5-fiuorouracil (group A), chemoembolization with Lipiodol (group B), and chemoembolization with Lipiodol and Gelfoam (group C). All patients were angiographically and sonographically followed up. RESULTS: The I-year and 3-year survival rates respectively, are 20.53% ± 3.62 and 1.95% ± 2.39 for group A, 51.34% ± 4.36 and 10.11% ± 4.92 for group B, and 62.99% ± 2.39 and 13.92% ± 4.96 for group C. The mean survival times for patients in each group were 9.57 months for group A, 16.07 months for group B, and 17.93 months for group C. CONCLUSION: Chemoembolization with Lipiodol and Gelfoam was the most effective therapy for latestage RCC in this study. 178 • 4:30 PM Phase II Trial of Hepatic Chemoembolization with Iodized Oil, Polyvinyl Alcohol, and CAM Chemotherapy M. C. Soulen, MD, Philadelphia, PA M.J. Pentecost, MD • L.A. Adams, RN R.D. Shlansky-Goldberg, MD • Z.J. Haskal, MD R.A. Baum, MD PURPOSE: We assessed the toxicity and response to chemoembolization with iodized oil and triple-drug chemotherapy. MATERIALS AND METHODS: Thirty-four patients (11 hepatoma, 19 colorectal, one cholangiocarcinoma, three other metastases) underwent 75 chemoembolizations with Ethiodol, polyvinyl alcohol, cis-platinum, doxorubicin, and mitomycin C. Response was determined with imaging studies and tumor markers. RESULTS: There were no complete responses. Patients with colon cancer had a 100% biological partial response rate (mean decrease in CEA level, 74%). CT scans showed minimal changes in tumor diameters. Among the 11 hepatomas, there were one partial and four minor morphologic responses and four were resected; two patients died without follow-up CT scans. From diagnosis, survival at 6, 12, and 24 months was 100%, 100%, and 73% for colon cancer and 80%, 60%,

1994 SCVIR Meeting Abstracts • 65 Volume 5 Number 1

and 60% for hepatoma, respectively. From first chemoembolization, survival at 6 and 12 months was 67% and 67% for colon cancer and 80% and 60% for hepatoma, respectively. Thirty-day procedural mortality was 4%. Major morbidity included lobar infarction (n = 2) and atrial fibrillation with myocardial ischemia (n = 1). Grade 3-4 toxicities included pain (25%), nausea/vomiting (10%), and anemia (1%). CONCLUSION: Chemoembolization with iodized oil and triple-drug therapy is reasonably well tolerated. Response rates and survival for hepatoma and colorectal cancer are substantially better than those achieved by systemic chemotherapy. 179 • 4:45PM Transarterial Chemoembolization for the Treatment of Primary and Secondary Neoplasms of the Liver: Long-term Follow-up R. Uflacker, MD, Charleston, SC PURPOSE: We assessed the efficacy of transarterial chemoembolization (TAC) in treating primary and secondary neoplasms of the liver when other treatments had failed. MATERIALS AND METHODS: TAC was performed in 70 patients, with a mean age of 54 years, who had hepatocellular carcinoma (HCC) (33 cases), metastatic breast carcinoma (MBC) (11 cases), metastatic colon carcinoma (MCC) (16 cases), and a variety oflesions (10 cases). A suspension of mitomycin C, Lipiodol, and Ivalon particles (300-600 !-lm) was injected into the hepatic artery in all patients every 6 weeks, up to seven procedures. CT and angiography were used for tumor response evaluation. Follow-up ranged from 3 to 87 months. RESULTS: Mean survival time was 13.9 months for the whole series. For patients with HCC, survival was 13.2 months; for MBC, 18.4 months; for MCC, 10.6 months; and for the remaining patients, 14.6 months. Improvement of clinical performance occurred in 83.6% of the patients. Tumoral response was observed in 75% of patients and progressive disease in 15%. CONCLUSION: TAC is an adequate alternative for palliation in patients with nonoperable primary and secondary liver neoplasms, resulting in clinical improvement, adequate tumor response, and increased survival time in most of the patients. TAC should be considered as a primary form of palliative treatment in patients with liver neoplasms.

180 • 5:00 PM Average Survival Period of Patients with Multiple Hepatocellular Carcinomas Treated with Chemoembolization and/ or Arterial Infusion Chemotherapy M. Honda, MD, Tokyo, Japan • T. Hashimoto, MD S. Matsui, MD • T. Gokan, MD H. Munechika, MD • T. Hishida, MD PURPOSE: We evaluated the average survival period of groups of patients with multiple hepatocellular carcinomas treated by means of different methods. MATERIALS AND METHODS: We divided 57 patients with multiple hepatocellular carcinomas into four groups as follows. Group 1, no treatment (n = 16); group 2, treated with chemoembolization (n = 14); group 3, treated with arterial chemotherapeutic infusion through implantable ports (n = 22), group 4, treated with arterial infusion chemotherapy after repeated chemoembolizations (n = 5). In the patients treated with chemoembolization a mixture of Adriamycin (60 mg) and Lipiodol (10 mL) was injected and gelatin sponge particles were delivered for embolization. In the patients treated with arterial infusion chemotherapy, adriamycin (40 mg/m 2 /4 W) and 5-fluorouracil (2,000 mg/m 2 /W) were injected. We compared the average survival period of each group. RESULTS: The average survival periods of group 1, 2,3, and 4 were 52,846,211 and 912 days, respectively. CONCLUSION: The patients with multiple hepatocellular carcinomas who could be treated with chemoembolization had longer average survival period.

Thursday Mternoon • 6D • Scientific Session 31 Abdominal Intervention MODERATOR:

Horacio D'Agostino, MD, San Diego,

CA 181 • 3:45 PM Percutaneous Drainage of Pancreatic Fluid Collections: Indications, Technique, and Results-Five-year Experience H.R. D'Agostino, MD, San Diego, CA R.M. O'Laoide, MB • S. Oglevie, MD B.F. Stainken, MD • B.F. Branstetter, MD E. vanSonnenberg, MD PURPOSE: We assessed the role of pancreatic drainage in the management of patients treated over a period of 5 years.

66 • Journal of Vascular and Interventional Radiology January-February 1994

MATERIALS AND METHODS: From July 1988 to July 1993, 34 patients (26 men, eight women) with pancreatic fluid collections underwent percutaneous catheter drainage at our institution. Fifteen patients had pancreatic pseudocysts, and 19 patients had infected pancreatic necrosis/abscess. The cause of pancreatitis included alcoholism (15 patients), postoperative (10 patients), gallstones (three patients), and miscellaneous (idiopathic, three patients; amyloidosis, one patient; intravenous lipids, one patient). Criteria for drainage were sepsis, symptoms (pain, discomfort, gastrointestinal or biliary obstruction), or enlargement in successive image studies. Percutaneous catheter insertion was the sole type of drainage in 18 patients (15 with pseudocyst, three with pancreatic necrosis, or three with abscess). Percutaneous drainage was combined with surgery in 16 patients (postoperatively in 15 patients, preoperatively in one patient). Criteria for drainage success was control of sepsis, relieffrom symptoms, and disappearance of the fluid collection. RESULTS: Success was achieved in 31 of 34 patients. Duration of drainage ranged from 12 to 279 days. Two patients had recurrence of pancreatic pseudocyst, and one patient had a residual pseudocyst after postoperative drainage of infected pancreatic necrosis. There were no procedure-related deaths. CONCLUSION: Pancreatic collections are effectively drained by percutaneous catheter placement. This approach plays a major role in managing pancreatic pseudocyst and infected pancreatic necrosis/abscess. 182 • 4:00 PM Chest Tubes: Physiology and Principles of Drainage of the Pleural Space M.A Braun, MD, Chicago, IL AA Nemcek, Jr, MD • R.L. Vogelzang, MD PURPOSE: Interventional radiologically placed smallcaliber chest tubes are a viable alternative to surgically placed large-bore chest tubes. Traditionally, thoracic surgeons have managed patients with chest tubes, and interventional radiologists have not been familiar with the principles of closed tube drainage (Pleur-Evac). Knowledge of basic principles can, however, allow the interventionalist to become comfortable with routine chest drainage. We assessed the utility of interventional radiologic management of pleural drainage. MATERIALS AND METHODS: Forty patients were treated for empyema (n = 26), hemothorax (n = 3), malignant pleural effusion (n = 6), and pneumothorax (n = 5). Pleural drains were left in for 1-30 days (average 5 days). The Pleur-Evac drainage system was assessed and adjusted daily by the interventional radiology service for air leak, tube function, and output.

RESULTS: In 25 of 26 patients with empyema and all of three with hemothorax, these complicated pleural collections resolved with use ofPleur-Evac drainage and aggressive use of thrombolytic agents and catheter repositioning. In six patients with malignant pleural effusion and three of four patients with a pneumothorax, evacuation of the pleural space collections was complete with use ofPleur-Evac drainage. CONCLUSION: Drainage of the pleural space can be entirely managed by an interventional radiology service without difficulty. 183 • 4:15 PM Outpatient Percutaneous Nephrostomy J.J. Snidow, MD, Indianapolis, IN V.J. Harris, MD • K. Kopecky, MD S.O. Trerotola, MD PURPOSE: This study was designed to test the hypothesis that percutaneous nephrostomy (PCN) can be safely performed as an outpatient procedure and to identify preprocedure patient variables that predict safety. MATERIALS AND METHODS: At the request of some referring physicians, a nontraditional outpatient approach to PCN evolved in our institution. Objective patient selection criteria have not been established, and thus patient selection has been made on the basis of individual referring physician preference tempered by individual radiologist's "judgment" regarding the suitability of each individual patient for outpatient management. A retrospective review of the medical charts was performed for all (n = 17) patients who underwent outpatient PCN in our practice between November 1991 and July 1993. An ongoing prospective study is currently in progress. RESULTS: Patient age ranged from 20 to 77 years (mean, 43 years). Both benign (n = 11) and malignant (n = 5) obstructions were treated. Procedure complexity included external PCN (n = 11), external nephroureterostomy (n = 5), and internal nephroureterostomy (n = 1) with tube size ranging from 8 to 12 F. No patients had coagulopathy, but five patients did have infected urine at the time of the procedure. None of the patients undergoing outpatient PCN required hospital admission for any periprocedure complications. CONCLUSION: PCN can be safely performed on an outpatient basis in selected patients. The number of patients studied thus far remains too small to identify preprocedure variables that predict safety.

1994 SCVIR Meeting Abstracts • 67 Volume 5 Number 1

184 • 4:30 PM Management of Cystic Necrotic Tumors: Role of Percutaneous Drainage and Sclerosis as Palliative Procedures to Control Symptoms and Sepsis H.R. D'Agostino, MD, San Diego, CA E. Morag, MD • R.M. O'Laoide, MB S. Oglevie, MD • B.F. Stainken, MD E. vanSonnenberg, MD PURPOSE: Our aim was to determine the role of percutaneous drainage and sclerosis on symptoms and sepsis in a series of patients with cystic and infected tumors. MATERIALS AND METHODS: From July 1988 to July 1993, 15 patients (nine women, six men) with cystic or necrotic tumors underwent percutaneous drainage at our institution. Tumors included cervical carcinoma (five patients), ovarian carcinoma (three patients), colorectal carcinoma (two patients), metastatic lung carcinoma (one patient), uterine sarcoma (one patient), bladder carcinoma (one patient), nasopharyngeal carcinoma (one patient), and cloacogenic carcinoma (one patient). Criteria for drainage were symptoms (pain, discomfort, or gastrointestinal, biliary, or urinary obstruction), or sepsis. Five patients underwent sclerosis with absolute alcohol. Criteria for drainage success were relief from symptoms and sepsis control. RESULTS: Successful relief of sepsis or control of obstructive symptoms was achieved in 11 of 15 patients. This effect was temporary in six patients and definitive in five patients. Duration of catheterization ranged from 3 to 90 days (mean, 23 days). One patient developed catheter tract seeding. Catheter dysfunction occurred in three patients. Collections that underwent alcohol ablation had the same evolution as those treated with drainage alone. There were no procedure-related deaths. CONCLUSION: Percutaneous drainage was effective for palliative treatment of symptomatic cystic and necrotic tumors in 73% of the patients in this series. No distinct benefits were obtained with the addition of alcohol ablation to percutaneous drainage. 185 • 4:45 PM Radiographic and Sonographic Follow-up of Urethral Stenosis Treated with the Wallstent J.M. Llerena, MD, Minneapolis, MN V. Garcia-Medina, MD • J.D. Berna, MD M. Thomas, MD • J.L. Rico, MD WR. Castaneda-Zuniga, MD PURPOSE: We evaluated the results of urethral stenosis treated with metallic stents (Wallstent) and compared conventional urethrography with urethral sonography in the follow-up of these patients.

MATERIALS AND METHODS: Eighteen patients (mean age: 52 years) with recidivant anterior urethral stenosis were treated with Wallstents placed endoscopically. Conventional retrograde and voidinl? urethrograms obtained with use of contrast mate~lal, as well as retrograde and voiding sonograms obtamed with use of sterile saline, were acquired at 1, 3, 6, and 12 months after stent placement. RESULTS: Initial evaluation showed resolution of the urethral stenosis in all 18 patients. In the next 3-6 months, different degrees of reepitheliatization of the Wallstent were noted in all 18 cases. At 12 months 13 patients (72%) demonstrated prominent mucosal growth within the stent, however no significant restenosis or obstruction was noted. Both standard urethrography and sonography were useful in defining the stent's position, urethral lumen, and mucosal regrowth. No significant complications were noted. CONCLUSION: Treatment ofrecidivant urethral stenosis with Wallstent endoprostheses was effective and safe. Stent reepithelialization was common in this series, but not associated with significant restenosis. Conventional urethrography and urethral sonography were equally effective in the follow-up of our patients, with some advantages for sonography because of the lack of ionizing radiation and its cross-sectional capabilities. 186 • 5:00 PM Treatment of Recurrent Posterior Urethral Strictures with Expandable Metallic Stents R.B. Gujral, MD, Uttar Pradesh, India • S. Roy S.S. Baijal • R. V. Phadke • R. Kukreti • S. Gupta PURPOSE: We evaluated the role of expandable Z stents in the treatment of recurrent posterior urethral strictures. MATERIALS AND METHODS: Eight male patients with recurrent posterior urethral strictures were treated by means of this method. Preprocedure work-up included radiographic urethrography and urodynamic evaluation. The procedure consisted of balloon angioplasty followed by placement of two stents in tandem. RESULTS: Successful dilation and placement of stents was possible in all cases. The only procedural complication encountered was slight proximal migration of the stents in one patient. Immediate postprocedure urethrography and urodynamic evaluation showed significant improvement. Follow up ranges from 5 months to 1 year. Dysuria necessitated stent extraction in the patient with proximal stent migration. Two patients had near total occlusion of the stented portion by an exuberant fibrous reaction, which required DVIU in one patient and open urethroplasty in the second. The remaining five patients

68 • Journal of Vascular and Interventional Radiology January-February 1994

continue to show a satisfactory clinical response correlating with similar findings on urodynamic evaluation. CONCLUSION: Short-term response in patients with posterior urethral strictures treated by Z stents appears favorable.

Thursday Mternoon • 6E • Scientific Session 32 Biliary Interventions Richard D. Shlansky-Goldberg, MD, Philadelphia, PA


187 • 3:45 PM Biliary Manometric-Perfusion Test and Clinical Trial: Value as Predictors of Long-term Patency Following Stent Placement for Primary Ductal Injuries or Biliary-Enteric Anastomoses

S.J. Savader, MD, Baltimore, MD A.C. Venbrux, MD • S.O. Trerotola, MD G.B. Lund, MD • S.E. Mitchell, MD F.A. Osterman, Jr, MD, et al PURPOSE: The purpose of this study was to evaluate the biliary manometric-perfusion test (BMPT) and "clinical trial" as predictors oflong-term patency following stent placement for primary ductal injuries or biliary-enteric anastomoses. MATERIALS AND METHODS: Forty-three patients with primary ductal injuries or biliary-enteric anastomoses underwent long-term biliary intubation (mean, 13 months) with 61 stents. Prior to removal of their stents, patients underwent a 2-3-week clinical trial (n = 35) with stents positioned above the treated region and in addition, a BMPT (n = 65) was performed to objectively evaluate the intrabiliary pressure under various perfusion rates. Patients were followed for 1-46 months after stent removal to determine the clinical outcome. RESULTS: The patients' clinical outcome was correctly predicted by the clinical trial in 23 of 26 (88%) patients and incorrectly predicted for three (12%) patients. The BMPT correctly predicted the clinical outcome in 37 of 43 (86%) patients and incorrectly predicted the outcome in six (14%) patients. KaplanMeier survival curve analysis of this data will also be presented. CONCLUSION: Both tests offer similar success rates predicting long-term patency in this group of patients.

188 • 4:00 PM Biliary Intervention via Percutaneous Puncture of Antecolic Jejunal Loop

L.J. Perry, MD, Boston, MA • K.R. Stokes, MD M.E. Clouse, MD PURPOSE: We evaluated the safety and efficacy of biliary intervention performed via percutaneous puncture of an antecolic jejunal loop in patients with surgical biliary-enteric anastomoses. MATERIALS AND METHODS: Thirty-five patients with antecolic Roux-en-Y or choledochojejunostomy/ hepaticojejunostomy loops underwent 46 interventional biliary procedures via direct puncture of the bowel loop. Indications for patients' previous surgery included sclerosing cholangitis (four patients), pancreatic cancer (one patient), orthotopic liver transplantation (16 patients), cholangiocarcinoma (four patients), and bypass of benign intrahepatic strictures (five patients). Procedures performed consisted of 22 ductal or anastomotic balloon dilations, six metallic and three plastic stent placements, two biliary drainages, one stone removal, one duct biopsy, and 14 diagnostic cholangiographic studies. RESULTS: This approach was successful in 43 of 46 attempts. In 19 patients, transhepatic puncture was entirely avoided. In one of the 35 patients, the loop was successfully catheterized, but further intervention from this approach was not possible because of unfavorable geometry. One major and two minor complications were encountered. No procedure-related deaths occurred. CONCLUSION: We report this approach as an effective and safe route for a variety of percutaneous biliary procedures in patients with biliary-enteric anastomoses. 189 • 4:15 PM Technique for the Prevention of Hemobilia When Placing Transhepatic Biliary Drainage Catheters

S.C. Goodwin, MD, Los Angeles, CA B.F. Stainken, MD • T.O. McNamara, MD PURPOSE: We evaluated a technique for the prevention of hemobilia when transhepatic biliary drainage catheters are placed. MATERIALS AND METHODS: Twenty patients with strictures were randomized to two groups. In the control group the biliary tree was accessed with an Accustick system and a transhepatic biliary drainage catheter was placed routinely. In the experimental group, following initial access, a rotating hemostatic valve was attached and the outer sheath was pulled back over the wire while contrast material was injected. If a major vascular structure was encountered, the tract was not used for transhepatic biliary drain-

1994 SCVIR Meeting Abstracts • 69 Volume 5 Number 1

age catheter placement. However, the outer sheath was readvanced and utilized to opacify the ducts. This facilitated separate access. Once access was achieved without traversing a major vascular structure, a transhepatic biliary drainage catheter was placed. Both groups were evaluated for hemobilia for 72 hours. RESULTS: There were three cases of hemobiIia in the control group; two were mild, one was severe. In the experimental group the tract communicated with a major vascular structure in three patients. In these patients a second access was utilized for placement of the transhepatic biliary drainage catheter. None of the patients in the experimental group experienced hemobilia. CONCLUSION: Visualization of the tract prevents the usage of tracts that communicate with large vascular structures. This reduces the incidence of hemobilia when transhepatic biliary catheters are placed. 190 •


Interventional Radiologic Management of Major Bile Duct Injuries Secondary to Laparoscopic Cholecystectomy S.L. Kaufman, MD, Atlanta, GA • L. G. Martin, MD AM. Zuckerman, MD • C.A Lewis, MD PURPOSE: We report our experience in the interventional radiologic evaluation and management of bile duct complications oflaparoscopic cholecystectomy. MATERIALS AND METHODS: Twelve patients underwent percutaneous biliary drainage for biliary complications after laparoscopic cholecystectomy. Six had initially undergone bilio-enteric bypass procedures after recognition of bile duct injury. Patients presented with jaundice, fever, right upper quadrant pain, and bilio-cutaneous fistulas. RESULTS: Cholangiography showed six patients with strictures at bilio-enteric anastomoses four with primary bile duct strictures, and two with bile ducts obstructed by surgical clips. Bile duct leaks were also seen in two patients. The highest level of biliary injury was the common hepatic duct in seven cases, the hepatic duct bifurcation in three, and above the hepatic duct bifurcation in two. Six patients underwent surgical repair. The other six underwent balloon dilation and large-bore catheter stent placement for strictures. Four patients are asymptomatic 5, 10, 16, and 18 months after stent tube removal. Stents are still in place in two patients. CONCLUSION: Percutaneous techniques may be used successfully in many patients with complications oflaparoscopic cholecystectomy. Given the cephalic extent of bile duct injury and occurrence of anastomo~ic strictures after bilio-enteric bypass, biliary dramage and stricture dilation should be given consideration as the initial, primary form of treatment.


4:45 PM

Percutaneous Intervention for Bile Duct Injury Following Laparoscopic Cholecystectomy C. Reiheld, MD, Charleston, SC • C.L. Sistrom, MD R. Uflacker, MD • 1. Vujic, MD PURPOSE: The rapidly increasing popularity of laparoscopic surgery has resulted in a concomitant increase of bile duct injuries in postcholecystectomy patients. Such cases often present difficult problems involving strictured, ligated, and discontinuous ducts as well as bile leaks and chronic bilomas. We present our recent experience with percutaneous treatment of these patients. MATERIALS AND METHODS: We have evaluated and treated 10 patients with bile duct injury following laparoscopic cholecystectomy. Using percutaneous biliary access as well as manipulation through indwelling T tubes and previously placed biloma drains, we have sought to achieve internal biliary drainage. A variety of devices were used including plastic stents, transhepatic biliary catheters, and wire mesh stents. Temporary bilateral transhepatic biliary catheterization was employed in two cases. RESULTS: In all cases adequate long-term biliary drainage was obtained. Bilirubin levels were normalized, bile leaks were controlled, and cholangitis resolved. No patients have required laparotomy for biliary repair or drainage. CONCLUSION: Bile duct injury following laparoscopic cholecystectomy presents an increasing challenge to interventionalists. We present several innovative techniques to deal with these often complex problems. 192


Use of Articulated Catheters for the Temporary Drainage of High Complex Biliary Lesions, Pancreatic Fistulas, and Difficult T-Tube Tracts R.D. Shlansky-Goldberg, MD, Philadelphia, PA C.A Chiaramonte, MD • C. Cope, MD M.C. Soulen, MD • R.A. Baum, MD Z.J. Haskal, MD, et al PURPOSE: Complex biliary obstructions and pancreatic fistulas are difficult to drain, usually requiring multiple catheters often placed through a single tract. Metal stents are often used for their ability to be positioned in multiple bile ducts through one tract. We have utilized a single articulated catheter when permanent stent placement was not desirable due to benign or treatable disease, and to replace surgical T tubes with difficult tracts. We investigated the indications, effectiveness, and follow-up of articulated draina.ge catheters placed to treat a series of biliary obstructIons, T-tube replacements, and pancreatic fistulas.

70 • Journal of Vascular and Interventional Radiology January-February 1994

MATERIALS AND METHODS: Articulated drains (8-14 F) made from segments of biliary catheters were placed in 15 patients from 1989 to 1993. Five were placed to drain multifocal or high biliary obstructions, two for pancreatic fistulas, and eight to replace surgical T tubes. Outcome was assessed with follow-up cholangiography and by means of ease ofremoval. RESULTS: Successful initial deployment occurred in all 15 cases, resulting in adequate drainage. There

were three occlusions (14-90 days) and one difficult removal. New articulated tubes were placed in all three patients. CONCLUSION: Articulated tubes represent an alternative therapy for complex lesions in both biliary and pancreatic ducts when permanent metal stents may not be initially desirable.