Surface Epithelium Mitoses Are a Strong Predictor of Cancer Progression in Barrett's Esophagus
DP Coco, A Srivastava, CA Sanchez, DS Cowan, BJ Reid, PL Blount, RD Odze. Brigham & Women's Hospital, Boston; Dartmouth Hitchcock Medical Center, Lebanon; Fred Hutchinson Cancer Center & University of Washington, Seattle
Background: At present, dysplasia is the clinical standard for risk stratification for progression to cancer in patients with Barrett's esophagus (BE). However, dysplasia interpretation suffers from a high degree of interobserver variability, and only a small proportion of BE patients develop esophageal adenocarcinoma (EA). Cell cycle abnormalities, especially those affecting G2 and mitosis (increased 4N), have been associated with EA in BE. In a previous pilot study, we reported an association between the presence of surface epithelium mitoses and progression to adenocarcinoma in BE. The aim of this study was to evaluate the prognostic significance of surface mitoses in a large prospective cohort of BE patients with long-term follow-up.
Design: 3999 routinely processed mucosal biopsies from 214 BE patients (M/F ratio:170/44, mean age: 63 years, mean BE segment length: 5.7 cm) followed for a mean of 90.4 months (range: 2.3-176 months), all of whom had a baseline (“index”) endoscopy between 1995 and 1999 and at least one follow-up endoscopy, were included. The development of adenocarcinoma was the primary outcome. All biopsies were evaluated in a blinded fashion for the number and percent of typical and atypical mitoses, in the surface and crypt epithelium, in both dysplastic and non-dysplastic epithelium. Data were analyzed using a Cox regression model to account for follow-up intervals and censored data.
Results: A strong positive correlation was noted between the presence of surface epithelium mitoses, both typical [p=0.0005, Hazard ratio (HR): 4.5, confidence interval (CI) 1.9-10.6] and atypical (p=<0.0001, HR: 7.9, CI 3.8-16.8) and the development of adenocarcinoma. In contrast, typical crypt mitoses showed no correlation with progression to cancer. However, the presence of atypical crypt mitoses showed a strong correlation with cancer outcome (p=0.0046, HR: 2.8, CI 1.4-5.9). All of these data were statistically significant regardless of the presence or absence of dysplasia.
Conclusions: The presence of surface epithelium mitoses, both typical and atypical, are a highly significant morphologic biomarker of cancer progression in BE. This is consistent with the previously well-known association between cell cycle abnormalities affecting the G2/M intervals and cancer progression in BE.
Wednesday, March 24, 2010 9:30 AM
Poster Session V # 65, Wednesday Morning