simplify the surveillance strategy. We propose a practical approach based on a novel optical technology - Spatial-domain Low-coherence Quantitative Phase Microscopy (SL-QPM) that utilizes the ultra-sensitivity of light interference effect and allows quantitative assessment of subtle structural alterations of the cell nuclei with a sensitivity of less than a nanometer. We hypothesize that a “field effect” associated with columnar epithelial cells from normal-appearing gastric cardia would exhibit subtle structural changes that are detectable by SL-QPM and allow the discrimination between those high-risk patients with EAC and dysplasia and those low-risk patients without. Methods: A pilot study was performed with 54 patients undergoing scheduled upper endoscopy. The patients were categorized into two groups: low-risk (14 patients; 5 Barrett’s intestinal metaplasia, 9 normal esophagus) and high-risk (40 patients; 16 EAC, 18 highgrade dysplasia, 6 low-grade dysplasia). We obtained the columnar epithelial cells via brushing from gastric cardia located at ⬃2 cm below the gastroesophageal junction. The cells were fixed with Cytolyt and subsequently prepared with Thinprep processor. Optical analysis was done by observers blinded to clinical/endoscopic data. Results: Several SL-QPM-derived nanoscale nuclear architecture parameters (standard deviation of optical path length, entropy, uniformity) were significant in discriminating high-risk from low-risk patients. A prediction model was developed based on logistic regression by combining SL-QPM-derived nuclear architecture parameters. We were able to accurately discriminate patients with neoplasia in 36 out of 40 patients (90% sensitivity) while correctly identifying 7 of 14 low-risk patients (50% specificity). Conclusion: The use of optical markers appears feasible in discriminating a highrisk group with esophageal neoplasia through the assessment of normal appearing cells from the gastric cardia, thereby identifying a high-risk patient population that warrants more intensive endoscopic surveillance. Future studies including expansion of patient populations and identification of additional optical or molecular markers to improve the prediction accuracy are necessary.
Sa1538 Endocytoscopy in Barrett’s Esophagus: Inter-Observer Variation Study Yutaka Tomizawa1, Kenneth K. Wang2, Ganapathy A. Prasad2, Louis-Michel Wong Kee Song2, Navtej Buttar2, Lori S. Lutzke2 1 University of Pittsburgh Medical Center, Pittsburgh, PA; 2Mayo Clinic Rochester, Rochester, MN Background: Endocytoscopy (ECS) has been applied to detect lesions in the gastrointestinal tract and has the potential to assess histological changes. We previously proposed a clinically relevant simplified classification system in Barrett’s esophagus (BE). The interobserver variability is critical in accurate diagnosis as interpretations of cellular and nuclear changes may be subject to the observer as histology. Aim: To assess the interobserver variability in endocytoscopic images in BE using our classification system. Method: Patients undergoing endoscopic mucosal resection were assessed ex-vivo using ECS. Lesions targeted for resection were endoscopically apparent areas of dysplasia or cancer in patients with known high grade dysplasia. Staining of the surface was done with 20% acetylcysteine and 1% methylene blue. Each specimen was imaged with a flexible catheter-type contact video endoscope (EndoCytoscope XEC 120, Olympus Tokyo Japan) with magnification x1100. These images were recorded for analysis and comparison. We previously proposed a modified ECA (endocytoscopic atypia) classification of four categories (mECA). The mECA classifications included cytoplasm-rich, rhomboid cells in a regular pattern (mECA1), increased cell number and different-sized nuclei/cells (mECA2), increased nucleus-cytoplasm ratio and dense chromatin and nucleus fission are prominent (mECA3), cells of various sizes are arranged irregularly, with blurred and enlarged nuclei (mECA4), and each classification is identical to squamous epithelium in esophagus, non-dysplastic BE, dysplastic BE, and esophageal adenocarcinoma, respectively. ECS naive gastroenterology fellows and faculty physicians in BE unit participated. After receiving a didactic session about ECS and viewing an atlas about each category, we showed randomly assigned unknown-diagnosis pictures and asked the observers to classify the images. Result: A total of 22 clinical fellows and 3 faculty physicians participated in this study. Using our classification system, overall accuracy for correct diagnosis in fellows and faculty physicians were 81.8% and 91.7%, respectively. Accuracies for each category are 97.7%, 100% for mECA1, 63.6%, 100% for mECA2, 70.5%, 83.3% for mECA3, and 95.5%, 83.3% for mECA4, respectively. If we combine mECA2 and mECA3 as diagnosis of BE using this system, the accuracy would be nearly perfect (98.9%) even in ECS naive observers. Conclusion: Overall, we had an excellent accuracy for diagnosis of squamous epithelium and esophageal adenocarcinoma and BE in both groups. Our classification system could be potential in endoscopic diagnosis of dysplasia in BE for gastrointestinal endoscopists. All misdiagnosis of mECA3 as dysplasia was answered as nondysplasia. This fact probably reflect some of the dilemma that exists with pathological interpretation of dysplasia in BE.
Sa1539 Long-Term Outcomes of Endoscopic Resection for Superficial Esophageal Neoplasms: Comparison Between EMR and ESD Method Do Hoon Kim, Hwoon-Yong Jung, Kee Don Choi, Ho June Song, Gin Hyug Lee, Kwi-Sook Choi, Jeong Hoon Lee, Mi Young Kim, Ji Yong Ahn, Ji Young Choi, Jin Ho Kim Asan Medical Center, Seoul, Republic of Korea Introduction: Endoscopic resection (ER) is well accepted method for treatment of differentiated early gastric cancer confined to mucosa without ulceration. For premalignant and early malignant lesions of the esophagus, ER can be applied instead of conventional esophagectomy for differentiated mucosal cancers. Methods of ER can be divided into endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD). Method: From December 1996 to December 2009, consecutive 137 superficial esophageal neoplasms in 122 patients were treated by ER. The indicated lesions were diagnosed as high grade dysplasia, squamous cell carcinoma or adenocarcinoma by endoscopic biopsy. Rates of en bloc resection, complete resection, and complications were evaluated as short-term outcomes. Overall survival, local or distant recurrence, and postoperative stricture were evaluated as long-term outcomes. Additional esophagectomy or chemoradiotherapy(CRT)/radiation therapy(RT) was warranted who had incomplete resection or submucosal invasion after initial ER. We analyzed the patients with SEC taken ER for initial treatment for evaluating the efficacy of two methods. Results: M:F ratio was 117:5 and median age was 66 (range, 45⬃84). 1) EMR group (37 lesions, 31 patients, M:F⫽31:0, median age⫽66). En Bloc resection was achieved in only 15 (40.5%) lesions. Complete resection rate was 70.3%. Stricture was developed in 3 patients (8.1%). Additional therapy was performed in 6 patients (surgery 2, CRT/RT 3, EMR 1). 7 patients died due to esophageal cancer or other disease. During median follow-up duration of 35.7 months, overall survival rate and disease free survival rate were 90.2 and 86.8% at 5 years. 2) ESD group (100 lesions, 91 patients, M:F⫽86:5, median age⫽66). En Bloc resection was achieved in 85 (87.6%) lesions. Complete resection rate was 92%. Stricture was developed in 4 patients (4.0%) and bleeding in one (1.0%). Ten (10.0%) perforated patient was recovered by clipping during ESD and conservative treatment. Additional therapy was performed in 7 patients (surgery 5, CRT/RT 2). During median follow-up duration of 16.8 months, overall survival rate and disease free survival rate were 100 and 98.9 at 4 years. Conclusion: ESD method may be an alternative curative measure for SEC in terms of safety as well as curability compared to EMR method.
Sa1540 Feasibility of Endoscopic Resection in Superficial Esophageal Cancer Ji Young Choi1, Young Soo Park2, Hwoon-Yong Jung1, Ji Yong Ahn1, Mi-Young Kim1, Jeong Hoon Lee1, Kwi-Sook Choi1, Do Hoon Kim1, Kee Don Choi1, Ho June Song1, Gin Hyug Lee1, Kyung-Ja Cho2, Jin-Ho Kim1 1 Department of Gastroenterology, University of Ulsan College of Medicine, Asan Medical Center, Asan Digestive Disease Research Institute, Seoul, Republic of Korea; 2Department of Pathology, University of Ulsan College of Medicine, Asan Medical Center, Asan Digestive Disease Research Institute, Seoul, Republic of Korea Introduction: Endoscopic resection in patients with superficial esophageal cancer (SEC) is limited by the presence of lymph node metastasis (LNM), highlighting the importance of determining which patients have virtually no risk of LNM. The objective of this study was to investigate the clinicopathological parameters predicting LNM in patients who underwent esophagectomy for SECs and to identify the best candidate patients for endoscopic resection. Methods: From 1991 to 2009, a total of 194 patients underwent esophagectomy for SECs in our institution. We evaluated the clinicopathological features associated with LNM. Results: Of 194 patients, 41 (21.1%) had LNM. The rates of LNM in patients with m1, m2, m3, sm1, sm2, and sm3 lesions were 0.0% (0/18), 8.5% (4/47), 25.0% (6/24), 15.0% (3/20), 27.6% (8/29), and 38.5% (20/52), respectively. On multivariate analysis, lymphovascular invasion (LVI) (p ⬍ 0.001), superficial tumor size (p ⫽ 0.002), and lower LMM (lamina muscularis mucosa) invasion width (p ⬍ 0.001) were independent predictors of LNM in patients with SEC invading the LMM. Among 64 patients with mucosal or sm1 cancer ⱕ 3 cm, only one had LNM without LVI showing lower LMM invasion width ⬎ 3.0 mm. Conclusions: Endoscopic resection should be carried out for mucosal cancer with superficial size ⱕ 3 cm and without positive lymph nodes. Moreover, if pathologic examination of the endoscopically resected specimens shows invasion of the sm1 layer and of ⱕ 3.0 mm lower LMM invasion width, indicating an absence of LVI, the patient can be carefully observed without additional treatment.
Volume 73, No. 4S : 2011