Tu1544 The Utility of Endoscopic Submucosal Dissection (ESD) for Rectal Neoplasms and the Clinical Short-Term Outcomes

Tu1543 Predictive Risk Factors Associated With Submucosal Fibrosis in Colorectal Endoscopic Submucosal Dissection Koji Takahashi*, Toshio Uraoka, Osamu Goto, A.I. Fujimoto, Yasutoshi Ochiai, Joichiro Horii, Naohisa Yahagi Division of Research and Development for Minimally Invasive Treatment, Cancer Center, Keio University School of Medicine, Tokyo, Japan Objective: Existence of fibrosis into the submucosal (SM) layer has been known as technical difficulty factor during colorectal endoscopic submucosal dissection (CRESD), however, predictive risk factor of SM fibrosis is still unclear. The aim of this study was to clarify the predictive risk factors of pre-, periand postCR-ESD associated with SM fibrosis. Methods: A total of 163 patients with 173 tumors underwent CR-ESD at our division from May 2010 to July 2013 were enrolled in this retrospective cohort study. The degree of SM fibrosis was classified into four types; Grade 0, no fibrosis; Grade 1, mild fibrosis in which blood vessels can be recognized; Grade 2, moderate fibrosis in which blood vessels can’t be recognized but indigocarmine dye added into SM injection solution can; Grade 3, severe fibrosis in which indigocarmine can’t be recognized). (a)Pre-operative risk factors: age, gender, tumor location, tumor size, macroscopic type, endoscopic diagnosis for SM deep invasion (cSM-d), performing previous biopsy. (b)Periand post-operative factors: procedure time, piecemeal resection, perforation, delayed bleeding, histological diagnosis for SM deep invasion (pSM-d). We analyzed these risk factors by chi-square test, Student’s t-test and logistic regression. Results: The mean tumor size was 34 19 mm. The median procedure time was 50 minutes (range 5-250). The rate of en bloc resection was 98%. Perforation and delayed bleeding occurred in 4 and 4 cases (2.3%), respectively. SM fibrosis grade distribution was as follows: Grade 0, 71 tumors (41%); Grade 1, 54 tumors (31%); Grade 2, 35 tumors (20%); Grade 3, 13 tumors (8%). (a) According to univariate analysis, there were significant differences in local recurrence (P!0.001) and cSM-d (p!0.001). When local recurrence lesions were excluded, there were significant differences in previous biopsy (pZ0.007) and non-granular-type laterally spreading tumors (LST-NG) (pZ0.037). In addition, when dividing of Grade into two groups in two ways (group 0-1 vs. 2-3 or group 0-2 vs. 3), there were significant differences in local recurrence, previous biopsy and LST-NG only in the case of group 0-1 vs. 2-3. Multivariate analysis showed that cSM-d (ORZ34.82, p!0.0001) and previous biopsy (ORZ4.01, pZ0.001) were independent risk factors. (b) According to univariate analysis, procedure time (p!0.0001) and pSM-d (pZ0.008) were significant factors, and even when Grade was divided, there was a significant difference in long procedure time (O50 minutes). Multivariate analysis showed that procedure time (ORZ2.70, pZ0.0064) was an independent factor. Conclusions: Based on this study, pre-operative diagnosis including estimated invasive depth was essential for determining severe submucosal fibrosis resulting in long procedure time. Previous biopsy should avoid in CR-ESD candidate case.