Background: Colorectal cancer (CRC) surveillance guidelines suggest that timing of a 3rd colonoscopy should be based on results of two prior exams. However, data are limited on whether baseline screening colonoscopy can inform the risk of advanced neoplasia (AN) at 3rd exam.
Methods: This study describes the risk of AN at 3rd colonoscopy stratified by findings on two previous exams in a prospective screening cohort and compares this risk over time from a negative 2nd exam between those with differing 1st exam findings.
The CSP #380 cohort included 3,121 Veterans aged 50-75 years who underwent screening colonoscopy from 1994-1997 and were followed for at least 10 years. Exclusion criteria included not having three colonoscopies more than one year apart, or having CRC at 1st or 2nd exam. The primary outcome was the proportion of AN at 3rd exam. Findings at 1st and 2nd exam were classified as high-risk adenoma (HRA), low-risk adenoma (LRA), or no adenoma. Chi-square tests compared proportions of AN on the 3rd exam between those with different baseline screening results but similar 2nd exam findings.
Results: This analysis included 891 participants: 58 (6.5%) had AN at 3rd exam. The proportion of AN at 3rd exam ranged from 3.2% to 21.4% when stratified by results of two previous exams. In participants with HRA or LRA on the 2nd exam, baseline screening colonoscopy was not associated with risk of AN at 3rd exam. However, for participants with no adenomas on the 2nd exam, baseline screening colonoscopy was associated with risk of AN at 3rd exam (P =.04). Furthermore, all AN was identified within about 5 years of the negative 2nd exam in those with HRA on the 1st exam.
Conclusions: Results of the 1st exam remain a risk factor for AN at 3rd exam in those with no adenomas at 2nd exam. This supports current guidelines which recommend a shortened surveillance interval in those with no adenomas at 2nd exam but HRA at 1st. Future work will combine CRC risk factors with genomic risk and colonoscopy outcomes over time to better identify individuals who might benefit from continued surveillance and to help inform appropriate surveillance intervals.
