Tu1138 NON-INVASIVE COLORECTAL NEOPLASMS REFERRED TO SURGERY: A PERFORMANCE KEY MEASURE FOR SCREENING PROGRAMS

Performance measures of endoscopic resection have not yet been defined. Second-look colonoscopy may reclassify neoplasms as superficial and amenable of endoscopic resection avoiding unnecessary surgery. Heterogeneity in operator competency and patchy distribution to advanced resection techniques can have unfavorable impacts on colorectal cancer screening. Observational, multicenter study including all consecutive colorectal neoplasms referred to surgery from all screening centers in a central Italy county within a 2-year period. Data were retrieved from county registry and prospectively collected charts from each center. Cancers were defined at endoscopy as superficial (Paris class) or deep (Borrmann class). Noninvasive cancer was definition if not invading submucosa (SM) on endoscopic or surgical specimens. Primary outcomes: noninvasive cancer rate according to endoscopic definition and resection (ER). Secondary outcome: indefinite cancer diagnosis at histology of ER specimens (lack of SM invasion reporting; indefinite microstaging of T1 cancer with free vertical-margin). 468 neoplasms from 13 centers were included. Patients: 64 y.o. (range 56-74), 256 males (58%). Neoplasms: superficial in 188 cases; deep in 280. Superficial neoplasms underwent: ER (Sup-ER) in 92 (49%); biopsies in 96 (Sup-B) (51%). Sup-ER were smaller (P<0.0001), more frequently pedunculated (P<0.0001), and in the left colon (P<0.0001) than Sup-B. A complete ER was achieved in 76 (83%) cases, which were smaller (P=0.006) and more frequently pedunculated (P=0.043). Overall noninvasive cancer rates of Sup-ER, Sup-B and Deep-B were 10%, 40% and 9%, respectively. Sup-B noninvasive cancer rate was significantly higher than Sup-ER (P<0.01); ulcer-negative Deep-B rate was significantly higher than ulcer-positive and stricturing Deep-B (P=0.0002). Indefinite cancer rate was significantly higher in Sup-ER with incomplete resection (P=0.012). Endoscopic center performance was significantly different for noninvasive cancer rate (range 0%-30%; P=0.0581), in particular for Sup-B (range 0%-100%; P=0.019). Noninvasive cancers rate was <10% in 5 centers and >20% in 5. Overall noninvasive cancer rate among patients referred to surgery in the present screening cohort is 15%, but it is widely heterogeneous due to significantly different rates in superficial neoplasms and deep neoplasms without invasive features (i.e. stricture or ulceration). Corrective actions should be: benchmark measures in endoscopic characterization; identification of referral centers both for 2nd-look colonoscopy and endoscopic resection were advanced techniques (i.e. ESD) should be available.