Findings in the distal colorectum are not associated with proximal advanced serrated lesions

Serrated lesions are an important contributor to colorectal cancer (CRC), notably in the proximal colon. Findings in the distal colorectum are markers of advanced proximal adenomatous neoplasia. However, it is not known whether they affect the odds of advanced proximal serrated lesions.

We performed a retrospective cross-sectional study of data from 1910 patients (59.3 ± 8.0 years, 53.8% female) who underwent an average-risk screening colonoscopy from August 2005 through April 2012 at Indiana University Hospital and an associated ambulatory surgery center. Colonoscopies were performed by an endoscopist with high rates of detection of adenomas and serrated polyps. Tissue samples of all serrated polyps (hyperplastic, sessile serrated adenoma/polyp [SSA/P], or traditional serrated adenoma) proximal to the sigmoid colon and serrated polyps >5 mm in the rectum or sigmoid colon were reviewed by a gastrointestinal pathologist and reclassified on the basis of World Health Organization criteria. Advanced serrated lesion (ASL) was defined as SSA/P with cytologic dysplasia, SSA/P ≥10 mm, or traditional serrated adenoma. Advanced conventional adenomatous neoplasia (ACN) was defined as tubular adenoma ≥10 mm, villous histology, high-grade dysplasia, or cancer. The prevalence of proximal ASL and ACN was calculated on the basis of distal colorectal findings. Multivariable logistic regression analysis was performed to determine the age-adjusted and sex-adjusted odds of advanced proximal adenomatous and serrated lesions. Secondary analyses were performed to examine the effect of variable ASL definitions.

Fifty-two patients (2.7%) had proximal ASL, and 99 (5.2%) had proximal ACN. Of the 52 patients with proximal ASL, 27 (52%) had no distal polyps. Of the 99 patients with proximal ACN, 40 (40%) had no distal polyps. Age and type of distal adenomas were significantly associated with proximal ACN. There were no significant associations between distal polyp type and proximal ASL. In secondary analyses, distal SSA/Ps (P = .008) but not distal hyperplastic polyps or conventional adenomas were associated with any proximal SSA/P.

The findings at flexible sigmoidoscopy that traditionally serve as indications for colonoscopy (conventional adenomas) are likely to be ineffective for detection of proximal ASL. This finding, plus the observation that most patients with proximal ASL have no distal polyps, favors screening colonoscopy over sigmoidoscopy, especially in the elderly. The observation that non-advanced distal SSA/Ps are associated with any proximal SSA/P warrants further study.