Vol 157. 1705-1710. May 1997 Printed i n U S A.
PROGNOSTIC FACTORS FOR RELAPSE IN STAGE I TESTICULAR SEMINOMA TREATED WITH SURVEILLANCE P. WARDE, M. K. GOSPODAROWICZ, D. BANERJEE, T. PANZARELLA, L. SUGAR, C. N. CA'ITON, J. F. G. STURGEON, M. MOORE AND M. A. S. JEWETT From the Drparftrrc't1t.s of Radiation Oncology. Medicine and Pathology, and Diuision of Urology, University nf Toronto and Toronto Hospital. and Drpcrrtn1rnt.s of Radiation Oncology, Medical Oncology. Surgical Oncology, Biostatistics and Oncolngic Pathology, Princess Margaret Hospital. Toronto, Ontario. Canada
Purpose: We sought to identify prognostic factors predictive of disease progression in patients with clinical stage I seminoma on surveillance following orchiectomy. Materials and Methods: Between J a n u a r y 1981 and December 1993, 201 patients 20 to 86 years old (median age 34) with clinical stage I seminoma were placed on surveillance following orchiectomy. The potential prognostic factors studied included age, tumor size, mitotic count, S phase fraction, ploidy, presence of small vessel invasion, syncytiotrophoblasts a n d tumor infilt r a t i n g lymphocytes, expression of @-humanchorionic gonadotropin a n d low molecular weight keratin on immunohistochemistry. Results: With a median followup of 6.1 years (range 1.3 to 12.3) 31 patients had relapse for an actuarial 5-year relapse-free rate of 84.9%. The 5-year actuarial survival rate was 97.1% and the cause specific survival rate was 99.5%. O n univariate analysis factors predictive of relapse were tumor size (5-year relapse-free rate 88 and 67% for tumors 6 cm. or less a n d greater than 6 cm., respectively, p = 0.0041, age (5-year relapse-free rate 79 a n d 91% for age 3 4 years or younger versus older than 34 years, respectively, p = 0.009) a n d presence of small vessel invasion (Byear relapse-free rate 86 versus 69%, p = 0.01). O n multivariate analysis age and tumor size were predictive of relapse, while small vessel invasion approached statistical significance. The risk of relapse in 57 patients with none of the 3 adverse prognostic factors (age greater than 34 years, tumor 6 cm. or smaller and no small vessel invasion) was 6%. Conclusions: We identified age, size of the primary tumor and small vessel invasion as important prognostic factors for relapse in patients with stage I seminoma treated with surveillance. Further followup and assessment of biological factors are needed t o optimize selection of patients at a high risk for relapse who should receive immediate postoperative therapy. KEY WORDS:seminoma, prognosis, testicular neoplasms
The standard treatment of patients with stage I testicular seminoma is inguinal orchiectomy followed by adjuvant radiation therapy to the para-aortic and ipsilateral lymph nodes.' While the acute morbidity of relatively low dose radiotherapy used in this setting has been minimal, there have been reports of impaired spermatogenesis, and increased rates of gastrointestinal symptoms and peptic ulceration on long-term follow~p.2-4Also, there has been increasing concern regarding the possible induction of second malignancies by radiotherapy.5." During the last 15 years there has been growing interest in surveillance of patients with stage I seminoma and reserving treatment for those with occult metastatic disease and relapse.7-10 The reasons for considering surveillance as a treatment option include the availability of computerized tomography (CT),which has allowed clinicians to stage disease more accurately and follow patients, and the development of curative chemotherapy regimens for those with relapse and extensive disease. The challenge for physicians involved in treating patients with stage I seminoma is to recommend the most appropriate therapy for individual patients. Clearly, with surveillance programs many patients can avoid unnecessary therapy but a proportion will have disease progression and require treatment. Therefore, it is important to study potential prognostic factors for occult metastatic disease to identify patients at high risk for relapse who should undergo adjuvant treatment Accepted for publication September 13, 1996.
postoperatively. In other seminoma surveillance studies tumor size and small vessel invasion have been identified as important prognostic factors for relapse.8.m Other histopathological factors, including invasion of the rete testis, histological subtype and necrosis, were of predictive value on univariate analysis but were correlated with tumor size and without independent prognostic significance on multivariate analysis. In patients with stage I seminoma treated with adjuvant radiation therapy other histopathological factors, including ploidy, have been identified as predictive for relapse, while in nonseminomatous germ cell tumors prognostic factors have been identified in many studies.11.12 Previously we identified age a t diagnosis as an important prognostic factor for relapse.'" We repeated a pathology review of the orchiectomy specimens, and present an update of our previously reported results with surveillance in stage I seminoma and an analysis of prognostic factors for relapse. MATERIALS A N D METHODS
Of 433 patients with stage I testicular seminoma treated between January 1981 and December 1993, 230 underwent our standard treatment consisting of radiation therapy to the retroperitoneal lymph nodes. Of these patients 203 were treated with surveillance, including 176 enrolled in a phase I1 prospective trial of surveillance that was begun in 1984. The major reason for declining entry into the surveillance study was patient preference for standard therapy. Two pa-
PROGNOSTIC FACTORS FOR RELAPSE IN SEMINOMA TREATED WITH SURVEILLANCE
tients treated with post-orchiectomy radiotherapy for stage I seminoma had a contralateral seminoma that was treated with surveillance. These patients were analyzed for outcome of both primary tumors. Additionally 25 patients not entered into the formal phase I1 study did not receive postoperative radiotherapy for various reasons and were observed outside the study. Two patients, 1 with concomitant advanced acquired immune deficiency syndrome and 1 with a nonseminomatous tumor on pathological review, are not included in this report. Thus, we present the results of surveillance following orchiectomy in 201 patients. Median age a t diagnosis of patients on surveillance was 34 years (range 20 to 86). Of the patients 20 (10%) had a history of cryptorchidism, including 11 who underwent orchiopexy. Radical inguinal orchiectomy was performed in 199 patients (99%)and scrota1 orchiectomy was done in 2. Postoperative evaluation included chest x-ray, CT of the abdomen and pelvis, and serum tumor marker estimation (a-fetoprotein [AFP] and b h u m a n chorionic gonadotropin [fl-HCGl). Followup evolved during the study. Currently patients are examined at 4-month intervals for the first 3 years, every 6 months for years 4 to 7 and annually from years 8 to 10. At each visit CT of the abdomen and pelvis is performed. Chest x-rays are done at alternate visits and serum tumor markers are estimated at each visit for the first 3 years of surveillance. Diagnosis of seminoma was confirmed by pathological review at our institution in all cases. For this analysis we obtained the paraffin blocks, when available, from the hospitals at which the orchiectomy was performed. Paraffin blocks were obtained for analysis in 139 cases and slides were available for review in an additional 15. All specimens were reviewed by 2 pathologists (D. B. and L. S.) without prior knowledge of the clinical outcome. Hematmylin and eosin stained paraffin sections were reviewed to confirm the diagnosis of testicdar seminoma; determine small vessel invasion, syncytiotrophoblasts, mitotic count per high power field and degree of tumor infiltration by lymphocytes, and select appropriate blocks for immunohistochemistry and deoxyribonucleic acid (DNA) analysis by flow cytometry. Immunohistochemistry for placental alkaline phosphatase, AFP, low molecular weight keratin and EHCG was performed using the avidin-biotin complex method for p a r a f i sections, with hematoxylin counterstain. Flow cytometric analysis of DNA content was performed using the technique described by Hedley et al.13 DNA histograms were analyzed using computer software with a rectangular model. Aneuploid peaks were defined as distinct peaks to the right of and separate from the W/G1 peak. Tetraploid peaks were defined as distinct peaks at the G2/M location containing greater than 20% of the total tumor nuclear preparation or a distinct peak to the right of the diploid G2h4 region. Single cut debris correction was used for S phase fraction calculations. The table outlines the number of patients with each attribute. Data on age and tumor size were available in all cases. However, because of our inability to obtain the paraffin blocks in all cases and technical difficulties with assessing variables in some (for example ploidy) the data were not complete. Complete pathological data were available in all cases with relapse on surveillance. Variables with skewed distributions (categories with less than 10 cases) included the expression of placental alkaline phosphatase and AFP on immunohistochemistry, and these were not examined for prognostic importance. The attributes assessed in the prognostic factor analysis included age, tumor size and histopathological factors, including mitotic count, S phase fraction, ploidy, presence of small vessel invasion, syncytiotrophoblasta. tumor infiltrating lymphocytes, and expression of PHCG and low molecular weight keratin on immunohistochemistry. The discriminant value for tumor size (6 cm.) was chosen based on the reoort from nnninh . . . . the _.__ ~ r . . .
Prognostic factors examined for relapse No. Pts.
Variable Age (yrs.): 34 or Younger Older than 34 Tumor size (cm.): 6 or Less More than 6 No. mitoses: Less than 3 3 or More Less than 5 5 or More DNA S phase %I: 29 or Less More than 29 Ploidy: Diploid Aneuploid Tetraploid Small vessel invasion: Absent Present Syncytiotrophoblasts: Absent Present Tumor infiltrating lymphocytes: Low/medium High
100 101 171 30 124 29 146 7 66 64 23 95 15 137 14 131 14 112 35
Pos. Low molecular wt. keratin: Neg. Pos. Placental alkaline phosphatase: Neg.
130 12 96 58
Testicular Carcinoma Study Group.8 The discriminant value for age was based on the median age of the study population. The survival and relapse-free rates were calculated from the date of diagnosis using the Kaplan-Meier meth0d.1~Differences in relapse-free rates with respect to individual baseline variables were calculated using the log-rank test.15 The Cox proportional hazards model was used to perform the multivariate analysis and assess the importance of variables defined on a continuous scale.16 RESULTS
With a median followup of 6.1 years (range 1.3 to 12.3) the &year overall actuarial survival for all 201 patients was 97.1% and the cause specific survival was 99.5%. Six patients died 1of recurrent seminoma and 5 of unrelated causes with no evidence of disease recurrence. Relapse occurred in 31 patients and the actuarial risk of relapse at 5 years was 15.1% (fig. 1).Relapse was observed 3 to 108 months after diagnosis (median 17). Of the 31 relapses 28 (90%)were in the para-aortic nodes and in 1there was concurrent involvement of the external iliac nodes. The median size of the para-aortic nodes at relapse was 3 cm. (range 1.5 to 9 cm.). One patient had concurrent relapse in the mediastinal and ipsilateral inguinal lymph nodes, 1 had a n isolated ipsilatera1 inguinal node relapse and 1had recurrence in the pelvic lymph nodes bilaterally. Biopsy confirmed the relapse in 10 of the 31 cases, and revealed seminoma in all instances. Of the 31 patients with disease progression 23 were treated with radiation therapy (22 to the para-aortic and pelvic lymph nodes, and 1to an isolated relapse in a right inguinal node). Six patients underwent chemotherapy and 2 (who had prior radiotherapy for an initial testicular seminoma) underwent retroperitoneal lymphadenectomy. All 23 patients treated with radiotherapy achieved in-field tumor control
PROGNOSTIC FACTORS FOR RELAPSE IN SEMINOMA TREATED WITH SURVEILLANCE
Years from orchiectomy
respectively (p = 0.009, fig. 3,A),and 86 versus 6% for 137 with no and 14 with capillary andor lymphatic invasion, respectively (p = 0.01,fig. 3,B). On multivariate analysis age and tumor size remained significant independent predictors of relapse, while small vessel invasion approached statistical significance. ARer adjustment for other prognostic factors, patients 34years old or younger were 3.6 times more likely to have relapse than those older than 34 years (95% confidence interval 1.4 to 9.2%. p = 0.004).Patients with primary tumors larger than 6 cm. were 2.8 times more likely to have relapse than those with tumors 6 em. or smaller (95% confidence interval 1.2 to 6.6%. p = 0.03).The relative risk of relapse for patienta with small vessel invasion was 2.3 (95% confidence interval 0.9 to 5.8%, p = 0.11).A total of 57 patients had no adverse prognostic factors (ageolder than34years, tumor 6 cm.or smaller and no small vessel invasion) and they had only a 6% risk of diaease progression.
FIG. 1. Relapse-free rate DISCUSSION
and 4 had a second relapse outside the irradiated volume 7to 20 months later: 2 in the mediaatinum alone, 1 in the lung, and 1 in the mediastinum and bone. These 4 patients with second relapse were treated with etoposide and cisplatin chemotherapy (3to 6 courses), and remained free of disease since completing treatment (21,42,51 and 88 months). Of 6 patients treated with cisplatin based chemotherapyat initial relapse 5 had bulky retroperitonealdisease (6to 9 cm.) and 1 had mediastinal disease. First and second line chemotherapy failed in the latter patient, who died of disease while undergoing high dose chemotherapy and autologous bone marrow transplantation. This patient represents the only disease related death on the surveillance protocol. Two patients, who previously received radiotherapy to the para-aortic area for a contralateral testis tumor and had progression while on surveillance for a second primary, underwent retroperitoneal lymphadenectomyand neither had a second relapse (2.6and 7 yeam postoperatively). On univariate analysis, when tumor size and age were analyzed as continuous variables, younger patients (p = 0.005)and those with larger tumors (p = 0.01)had a greater risk of relapse. In addition small vessel invasion in the primary tumor (p = 0.01) was of prognostic importance. No other factor reached conventional levele of significance. The relapse-free rate at 5 years was 88 versus 67% in 171 patients with tumors 6 cm. or smaller and 30 with tumors larger than 6 cm., respectively (p = 0.004, fig. 2),91 versus 79% in those older than 34 and 34 years old or younger,
Several prospective studies of surveillance in stage I testicular seminoma have been conducted during the last de-
cade.'-lo The largest reported experience is from the Danish Testicular Carcinoma Study Group with a crude relapse rate of 19% in 261 patients a t a median followup of 48 months.* On d v a r i a t e analysis tumor size, histological subtype, nerrosis and invasion of the rete testis were predictive of relapse. On multivariate analysis tumor size was the only signiscant factor with a cumulative risk of relapse after 4 years of 6% in those with tumors smaller than 3 cm.,18% in those with tumors 3 to less than 6 cm.and 36% with tumors 6 cm.or larger. In the Royal Marsden Hospital series of 103 patients treated with surveillance the 5-year relapsefree rate was 829b.10 The only significant factor predicting for relapse was lymphatic and/or vascular invasion (9 versus 17% relapse rate). We found tumor size, age and, to a lesser extent, small vessel invasion to be predictive of relapse. The prognostic importance of tumor size and small vessel invasion is not surprising given the findings in the Danish and Royal Marsden Hospital studies. However, age was not a prognostic factor in the l a e Danish surveillance stud9 and was not analyzed in the k y a l Marsden Hospital study.10 One could hypothesize possible reasons for the greater risk of relapse in younger patients. Age may reflect a more aggressive biology of disease in younger patients, as suggested by the pattern of age distribution for more aggressive noneeminomatousgerm cell tumors in younger patients, and progressively less aggressive pure seminomas and spermatocytic seminomas in older patients. However, age may also be a surrogate for other attributes not immediately clinically apparent. We assessed several other histopathological factors for prognostic importance, including DNA ploidy status, mitotic u count, DNA S phase per cent, syncytiotrophoblasts, degree of CI lymphocytic infiltration of the primary tumor, and expres2 sion of &HCG and low molecular weight keratin on immuW W nohistochemistry. No factor was associated with a signi642 cantly greater risk of relapse in these patients. Given the W v) small number of events in our study and the fact that data on 0.4 Q p=.OO4 all of these factors were not available in all cases, our conm u clusions regarding these factors must be interpreted with er caution and should be independently confirmed. Also the prevalence of certain potentially adverse factors, for example syncytiotrophoblasts, was low, which compromised evaluaof significance. It is possible that with further followup 0 2 4 6 8 1 0 1 2 1 4 tion and more events additional histopathological features or other factors will be recognized as predictive for oceult me& Years from orchiectomy 0 astatic disease. DNA ploidy status has been shown to be oh pmgm&k FIG. 2. Relapee-&ee mk in 171 patients with tumors 6 cm.or importance in clinical stage I nOPBBmiPOmBtOUB gemn esll mdler vereue 30 with tumors larger than 6 cm.
PROGNOSTIC FACTORS FOR RELAPSE IN SEMINOMA TREATED WITH SURVEILLANCE
p = .009
.______ >34 2
8 1 0 1 2 1 4
Years from orchiectomy
p = .01 Negative ........ Positive
1 0 1 2 1 4
Years from orchicctomy
FIG.3. Relapse-free rates in 98 patients 34 years old or younger versus 103 older than 34 years (A) and 137 without versus 14 with small vessel invasion (B).
testicular tumors.17 However, little data are available on its the patient and health care system associated with the use of usefulness in testicular seminoma. Mor et al reported on 65 routine postoperative radiation therapy. Also, while the patients with stages I and I1 seminoma treated with post- acute side effects of adjuvant radiotherapy are easily conorchiectomy radiotherapy or chemotherapy.18 Tumor pro- trolled with antiemetic and antidiarrheal medication, the gressed in only 5 patients with an aneuploid histogram. long-term risk of second cancers induced by radiation should Winkler et al reported on 97 patients with stages I to IV not be ignored.5.6 The most attractive feature of surveillance seminoma seen between 1960 and 1980 and found that none is the avoidance of unnecessary treatment in a large proporwith diploid tumors had progression compared to 13% with tion of patients. However, surveillance must not compromise aneuploid tumors.19 Ploidy was not assessed in the Danish8 survival and it is reassuring to find that there have been only or Royal Marsden Hospitallo studies of surveillance. While 4 disease related deaths in the 756 patients on surveillance our results indicate that ploidy status is not a discriminating in the Danish: Royal Marsden Hospitallo and our series. It is fador for relapse in stage I seminoma, further studies are important to emphasize that careful and prolonged followup necessary to confirm this finding. (minimum 10 years) is essential if these excellent results are There is considerable controversy in the literature regard- to be maintained. Physicians may view 1management option ing the prognostic importance of mitotic count in testicular as preferable but individual patient preferences must be seminoma. The World Health Organization criteria for diag- considered, which may be based on the side effects associated nosis of anaplastic seminoma (3 or more mitoses per high with both approaches. For example young patients in whom power field) have not been uniformly adopted, and most fertility is an issue may elect surveillance, while for others reported series indicate that the prognosis is similar for surveillance, with its inherently greater relapse risk, may be patients with anaplastic and classic seminoma. Indeed, some inconvenient and adjuvant treatment may be preferred. Also have suggested that anaplastic seminoma be abolished as a various socioeconomic factors could influence patient preferseparate clinical entity.20 In our series the mitotic rate was ences, for example desire to travel and consequent inability measured by 1pathologist in a uniform manner in all cases to be available for followup. and, when assessed as a continuous variable or using criteria It would be preferable to base recommendations regarding described in the literature to define anaplastic seminoma (3 treatment aRer orchiectomy on results from a randomized or more and 5 or more mitoses per high power field), it was phase 111 clinical trial comparing radiotherapy or surveilnot of value in predicting relapse. However, in patients with lance. Such a study would be difficult to conduct, since it stage I seminoma treated with adjuvant radiotherapy our experience12 and others21 indicate that anaplastic histology would have to be an equivalence study (primary end point of is associated with a considerably worse prognosis and, there- survival) with a low event rate. The available data suggest that almost 100% of patients with stage I seminoma are fore, further study is clearly necessary in this area. Several studies of other solid tumors have indicated that cured regardless of the approach chosen as post-orchiectomy the degree of tumor infiltration by lymphocytes and the per- therapy and any difference in outcome from second maligcentage of cells in the S phase are important determinants of nancies would not occur for many years. The identification of prognostic factors for relapse on surtumor aggressiveness.22 Also reports have indicated that veillance will allow clinicians to recommend the most approsyncyseminomas with immunohistochemically detectable tiotrophoblastic cells might represent a particular his- priate therapy for individual patients. As stated, the power of topathological entity with an unfavorable prognosis.23~24 our analysis to detect potentially clinically important progHowever, no factor appeared to affect outcome adversely in nostic factors was limited, and our conclusions must be our experience. The prognostic importance of focal cytokera- interpreted with caution and should be independently tin expression as detected on immunohistochemical staining confirmed. We anticipate that, with further knowledge was assessed, since its presence in testicular tumors is con- regarding folIowup and assessment of biological factors presidered to indicate a nonseminomatous germ cell malignan- dictive for occult metastatic disease, it may be possible to identify patients at low risk for relapse with much greater C Y . ~The ~ absence of placental leukocyte alkaline phosphatase is commonly seen in spermatocytic seminoma and, accuracy and recommend treatment to select patients only. therefore, classical seminoma lacking placental alkaline The principal investigators in the surveillance study were phosphatase expression could represent unrecognized sper- Dr. G. M. Thomas, Toronto Sunnybrook Regional Cancer matocytic seminoma.26 However, no marker was of value in Center (1984-1988); Dr. W. Duncan, Western General Hospredicting relapse in this series. pital, Edinburgh, Scotland (1988t o 1989); Dr. A. Munro, St. Bartholomew’s Hospital, London, England (1989 to 1990), CONCLUSIONS and Dr. M. Gospodarowicz, Princess Margaret Hospital, Surveillance is a viable alternative option in stage I sem- Toronto, Ontario, Canada. The current Drincipal investigator inoma. There are considerable social and economic costs to is Dr. P. Warde. I
PROGNOSTIC FACTORS FOR RELAPSE IN SEMINOMA TREATEI)WITH SURVEILLANCE REFERENCES
23. Bosman, F. T., Gierd, B W.,N ~ u w e n h w Kruseman, A. C.. Knijnenburg, G. and Spaauder, P.J.: Human chorionic gona1. Thomas, G., Jones, W.,VanOosterom,A. and Kawai, T.: Consendotrophin and alpha-fetoprotein in teaticular germ cell tusus statement on the investigation and management of M c m o m : a retmqective immunohietocbemioal study. Histopaular seminoma. Prog. Clin. Biol. Res., 351: 285,1990. thology, (1: 673,1980. 2. Hamilton, C.,Horwich, A, Easton, D. and Peckham, M. J.: Radiotherapy for stage I seminoma teetie: resultsof treatment 24. Butcher, D. N., Gregory,W.M., Gunter, P. A, Mastere,J. R.and Parkinson, M. C.: The biological and clinical eisnificance of and complications. Radiother. Oncol., 6: 115,1986. HCG-wntaining cells in seminoma. Brit. J. Cancer, 6k 473, 3. Hahn, E. W.,Feingold, S. M., Simpmn, L. and Batata, M.: 1985. Reeovery from aspermia induced by lowdose radiation in sem26. Eglen, D. E. and Ulbright, T. M.: The M e r e n t i d disrprosie of inoma patients. Cancer, M): 337,1982. yolk sac tumor and seminoma. Uaefulnees of cgtoberatin, 4. Yeoh, E., Horowitz, M., Rusm, A, Muecke, T., Robb, T. and alpha-fetoprotein and alpha-l-antitrgpein i m m u n ~ p e d d M e Chatterton, B.: The effecta of abdominal irradiation for seminoma of the testis on gastrointestinal function. J. Gastroreactions. h e r . J. Clin. Path., 8& 328,1987. Enterol. Hepatol., 10: 125,1995. 26. Aguirre, P., Sculley, R. E., Dayal, Y. and DeLek, R k. 5. Meller, H., Mellemgaard, A, Jaahen, G. K., Pedersen. D. and Placenta-like alkaline phosphatase in germ cell tumors ofthe Storm, H. H.: Incidence of second primary cancer following ovary and testis. Lab. Invest., 6% 2A.abetract, 1986. testicular cancer. Eur. J. Cancer, a9A:672,1993. 6. van Leuwen, F. E.,Stiggelhut, A M., van den BeltDueebout, EDITORIAL COMMENT A. W., Noyon, R., Eliel, M. R., van Kerkhoff, E., Delemarre, Since the early reporta of Peckham’ and Olive9 et al in 1983 the J. F. and Somers, R.: Second cancer risk following testicular cancer: a follow-up study of 1,909 patients. J. Clin. Oncol., 11: succeas of surveillance as a therapeutic modality for stage I semi415,1993. noma has been demonstrated in several other studies (mferences 8 7. Warde, P. R., Gospodamwicz, M. K., Goodman, P. J., Sturgeon, and 10 in article).a Even critics of this approach would probbly J. F., Jewett, M. A, Catton, C. N.,Richmond, H., Thomas, agree that the results of these studiea do not jeopardize the cure for G. M., Duncan, W. and Munro, A. J.: FZmults of a policy of patients with relapse from surveillance in this present era of excelsurveillance in stage I testicularseminoma. Int. J. Rad. Oncol. lent chemotherapy. Thedore, within the limitations of the current Biol. Phys., 27: 11, 1993. length of followup in these studies. it would seem that surveillance is 8. von der Maase. H., Specht, L., Jambsen, G. K., J a k o k ~ A, , therapeutically equivalent to retroperitoneal radiotherapy (referMadsen, E. L., Pedersen, M., Fbrth, M. and schultz, H.: Sur- ence 12 in article). Conversely. the proponenta of surveillance have veillance following orchiectomy for stage I seminoma of the selected a subgroup of patients who are compliant with an inteme testis. Eur.J. Cancer, -A: 1931,1993. and meticulous followup program. Thus. any deviation from such 9. Allhoff, E. P., Liedke, S., de Riese, W.,Stief, C. and Schneider. tight surveillance may conceivably alter the conclusions of these B.: Stage I seminoma of the teatie. w u v a n t radiotherapy or studies. Consequently, the ‘real world” application of surveillance surveillance? Brit. J. Urol., 68.190,1991. and success of this modality must follow as strict a followup regimen 10.Horwich, A., Alsaqiari, N., A’Hem, R, Nicholls, J., Dearnaley, as those reported if equivalent reaulta are to be expected. With D. P. and Fisher, C.: Surveillance following orchiectomy for stratification of patients by age (reference 12 in article) and tumor stage I testicular seminoma. Brit. J. Cancer, Wk 775,1992. eize (reference 8 in article) it may be poseible to customize the 11. Read, G.,Stenning, S. P., Cullen, M. H., Parkhmn, M. C., Horwich, A, Kaye, S. B. and cook,P. A: Medical Research intensity of the followup to the individual patient, with a musequent Council prospective study of surveillance for stage I te~ticular decrease in expenditurea and inconvenience to the patient, while teratoma. Medical Research Council Testicular Tumors Work- enhancing the surveillance or treating the high risk patients with radiotherapy. These maneuvers also have the potential to reault in ing Party. J. Clin. Oncol., 10: 1762,1992. 12. Warde, P., Gospodarowicz, M. K., Panzarella, T., Catton, C. N., better overall compliance of patients. In addition, it will be for the Sturgeon, J. F., Moore, M., Goodman, P. and Jewett, M. A: proponents of surveillance to argue convincingly why this modality Stage I testicular seminoma: results of a u v a n t irradiation should be used in view of the excellent resulta and low morbidity achieved with modem low dose (2,500cGy.) retroperitoneal radioand surveillance. J. Clin. Oncol., 13: 2255,1996. 13. Hedley, D.W.,Friedlander, M. L., Taylor, I. W..Rug& C. A. and therapy.‘ Since the therapeutic efficacy and morbidity of the 2 comMusgrove, E. A.: Method for analysis of cellular DNA content peting modalities are unlikely to be M e r e n t and patient preference of paraftin-embedded material using flow eytometry. J. Histo- m o t be predicted, then the remaining variable remains cost. In this day and age, when medical expenditures must be critically chem. Cytochem., 31: 1333,1983. 14. b p l a n , E. L. and Meier, P.: Nonparametric eatimation from justified, it is imperative that the costa of surveillance of patients with stage I seminoma be compared to those of adjuvant radiotherincomplete observations. J. h e r . Stat. h. 6% ,457,1958. 15. Mantel, N.: Evaluation of ~urvivaldata and two new rank order apy. Finally, the inclusion of markers predictive for relapse in any statistics arising in its mnsideration. Cancer Chemother. Rep.. algorithm determining the intensity of followup will likely minimiu! 50: 163,1966. the resultant cost for any surveillance protocol by focusing resources 16. COX,D.R.: Regression models and lifetables. J. Roy. Stat. h., on patients at highest risk for relapse while mSmizira those on the 34: --.) 187. --.-. 1972 majority of patients. Therefore cost will likely have a signiscant role 17. Austenfeld, M. S.,Bilhartz, D. L.. Nativ, 0.. Farrow, G. M. and in detewhether surveillance becomes a mainetream option Lieber, M. M.: Flow eytometric DNA ploidy pattem for pre- for patients with stage I seminoma. dicting metastasis of clinical stage I n~nsemin~matous gem Dan Theodorescu cell testicular tumors. Urology, 41: 379,1993. Division of Uml& onodogy 18. Mor, Y.. Leibovich, I., Raviv, G., Nass, D., M *, o., University 6f Viginia Cancer Center Goldwasser, B. and Nativ, 0.: Testicular seminoma: clinical Charlotte~vilk,Virginia sigdicance of nuclear deoxyrihnucleic acid ploidy pattem as studied by flow eytometry. J. Urol., 1W 1041,1995. 1. Peckham, M. J., Barrett, A, Horwich. A. and Hendry, W. F.: 19. Winkler, H. Z.,Nativ, O., H A , Y..Reiman,H. M. and Leber, Orchiectomy alone for stage I testicular non-seminoma. A M.M.: Testicular seminoma: nuclear DNA ploidy studied by pmqpvm report am the Rnyal Marsden Hospital study, Brit. flow eytometry. J. Urol., part 2,139.2064 abstract 174,1988. J. Urol., 58: 754,1983. 20. Hanks,G. E.,Peters, T. and Owen, J.: Seminoma of the testk 2. Oliver, R T., Hope-Stone, H. F. and Blandy, J. P.: Justification of long-term beneficial and deleterious results of radiation. Int. the w e of surveillance in the management of stage I germ cell J. Rad. Oncol. Biol. Phys., a4:913,1992. t u m o m of the testis. Brit. J. Urol., 6ik 760.1983. 21. Bobba,V. S.,Mittal, B. B., Hoover, S.V.and Kepka, A: Classical 3. Duchesne, G.M., Horwich, A, Dearnaley, D.P., Nicholls, J.. Jay, and a n a p l d c &oma: difference in survival. Radiology, G.. Peckham, M. J. and Hendry, W. F.:Orchidedomy alone for 161: 849,1988. 22. E b ,C. N., Bumette, J. J.. Sedlak, R,Dyas, C..yd B l e m o r e , stage I seminoma of the teatis. Cancer, 6& 1116,l~. W. S.:Prognostic applicatiw of DNA d p IDe~ d d W - 4. Thomas, G. M.: Surpeillanoe in stage I swnimma dthe t&b. Urol. Clin. N.h e r . , $0: 85,[email protected]
mt leeiom in humans. Surg..,Gynec & O W ,17%329,1991. ~
PROGNOSTIC FACTORS FOR RELAPSE IN SEMINOMA TREATED WITH SURVEILLANCE REPLY BY AUTHORS
Several valid points regarding current issues in the management of stage I seminoma are raised, particularly emphasis on the low morbidity of retroperitoneal radiation therapy. While this is correct, we should not forget that a proportion of patients have acute gastrointestinal toxicity, require antiemetic therapy and are unable to work for 3 to 4 weeks while on treatment. Also, the potential late toxicity of radiotherapy cannot be ignored. There is a slight but definite and well documented increased risk of a second malignancy in patients with seminoma treated with radiation therapy. This has been well documented in 2 large recent aeries with long-term followup (references 5 and 6 in article). Currently, a study of second cancers in testis cancer survivors is being performed by the Radiation Epidemiology Branch of the National Cancer Institute.
This increased risk of second tumor induction is especially important in patients with stage I seminoma, because more than 80%)ape cured with orchiectomy alone. Concerns regarding the cost of medical care should not be ignored. However, it is difficult to calculate accurately the overall cost of an increased risk of delayed potentidly lethal late toxicity, such a s a second malignancy. These issues must be considered when comparing the cost of prophylactic treatment versus surveillance. Available data suggest that almost 100%of patients with stage I seminoma are cured no matter what approach is chosen a s postorchiectomy management. We anticipate that with further knowl. edge of the factors predictive of occult metastatic disease and patterns of failure it should be possible to tailor management recommendations for individuals.