[587] Cancer Risk Stratification in UC Patients with Endoscopically Resected Dysplastic Polyps: Top-Down vs. Bottom-Up Dysplasia

X Gui, TA Ullman, N Harpaz. GI Pathology, PLLC, Memphis, TN; Mount Sinai School of Medicine, New York, NY

Background: Dysplastic polyps in the setting of UC are currently classified endoscopically as non-adenomalike, i.e., incompletely resectable, poorly circumscribed, irregularly-shaped or with stigmata of cancer, versus adenomalike, i.e., resectable, well-circumscribed and dome-shaped. Patients with non-adenomalike dysplastic polyps are at high risk of colorectal cancer (CRC) and are usually managed with immediate colectomy, whereas those with adenomalike polyps may be managed with polypectomy and endoscopic surveillance. Although these polyps overlap histologically, distinguishing histological criteria would be desirable, especially since endoscopists often find the endoscopic distinctions difficult to apply in practice. We carried out a cohort study comparing UC patients with endoscopically resected dysplastic polyps featuring superficial or top-down (TD) dysplasia and full-thickness or bottom-up (BU) dysplasia with respect to progression rates to CRC.
Design: Cohorts identified from an institutional database of UC patients undergoing surveillance examination in 1996-7 included patients with at least 1 dysplastic polyp and patients who were dysplasia-free (NoD) at their first colonoscopy at our institution. Each polypectomy slide was reviewed retrospectively and classified as having TD or BU dysplasia. Incompletely resected and poorly oriented polyps were excluded. Patients were followed until the development of colorectal cancer or censored at their last cancer-free examination. Differences in actuarial progression were examined by log-rank testing.
Results: Of 76 UC patients with dysplastic polyps (median follow-up 3.0y), 37 had at least one TD polyp and 39 had only BU polyps. 1 patient from the TD group and 5 from the BU group ultimately progressed to CRC. This difference was actuarially significant (p=0.016 by Log Rank testing). There was no difference in progression between patients in the TD and NoD groups (N=349, median follow-up 7.0y), 18 of whom developed CRC (p=0.86).
Conclusions: The top-down pattern of polypoid dysplasia in UC is a histological marker of low risk for CRC development, permitting conservative management with polypectomy and continued surveillance. The bottom up pattern may encompass lesions of low and high risk, and requires evaluation based on the current endoscopic criteria.
Category: Gastrointestinal

Tuesday, March 10, 2009 9:15 AM

Platform Session: Section C, Tuesday Morning