Colorectal cancer after surveillance colonoscopy: false-negative examination or fast growth?

PURPOSE: Colonoscopy is the preferred method for colorectal cancer surveillance of high-risk patients. Despite its high sensitivity, polyps or cancers may be undetected by colonoscopy and later attributed to an accelerated adenoma-carcinoma sequence. This study assesses how the characteristics of colorectal cancers found at intervals between surveillance relate to the adenoma-carcinoma sequence and its prevention. METHODS: The records of 557 patients with colorectal cancer that were diagnosed from January 1, 1990, to December 31, 1996, were reviewed to identify those patients who had prior colonoscopic surveillance within 60 months of their diagnosis. RESULTS: There were 29 (5.2 percent) patients who had one or more colonoscopies before diagnosis of their colorectal cancer. Mean interval between diagnosis and prior colonoscopy was 23 (range, 4-59) months. The distribution of cancers included nine cecum, two ascending, three hepatic flexure, five transverse, one splenic flexure, three descending, two sigmoid, three rectum, and one anal canal. The mean tumor size was 4.4 cm for the cecum and 2.4 cm for all other locations. There were 7 Tis, 6 T1, 4 T2, and 12 T3 lesions. Six patients with T3 lesions had prior colonoscopies within 24 months of the diagnosis. Three of four patients with lymphatic metastases had tumors in the cecum. Twenty tumors (69 percent) were well or moderately differentiated. Mean follow-up was 41 (range, 7-95) months with two local recurrences and two unrelated deaths. CONCLUSIONS: Size, differentiation, and stage of colorectal cancer in addition to the interval to diagnosis suggest that the majority of cancers found during surveillance colonoscopy followed prior false-negative examinations. Because cecal landmarks are most constant, prior photographic documentation may help to prove or disprove fast growth of cancers found in the cecum during surveillance colonoscopy.