Poster Session I – Poster of Distinction - A93 THE LONGER THE WAIT, THE GREATER THE RISK: A META-ANALYSIS OF SURVEILLANCE INTERVALS FOR PATIENTS AFTER RESECTION OF 1-2 SMALL NON-ADVANCED ADENOMAS

TL;DR

A meta-analysis finds that longer surveillance intervals after removing small non-advanced adenomas increase the risk of colorectal cancer and advanced adenomas.

Contribution

This is the first meta-analysis comparing surveillance intervals directly within the same at-risk population of patients with 1-2 small non-advanced adenomas.

Findings

Longer surveillance intervals are associated with a dose-dependent increase in colorectal cancer (CRC) and advanced adenoma (AA) risks.

Surveillance at <5 years compared to 6-10 years significantly lowers CRC and AA risks.

The protective effect is strongest when comparing <10 years versus >10 years intervals for CRC and AA.

Abstract

The optimal post-polypectomy surveillance interval for patients with 1-2 small (<10 mm) non-advanced adenomas (NAAs) is uncertain. Previous guidelines were informed by systematic reviews and meta-analyses (SRMAs) that used inappropriate comparator populations (polyp-free individuals or general population), rather than directly comparing surveillance intervals within the same at-risk population.

This SRMA is the first to compare different surveillance intervals in adults with a baseline finding of 1-2 small NAAs.

This SRMA was conducted to inform the CAG post-polypectomy surveillance guideline. We searched MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials up to March 2025 for randomized controlled trials (RCTs) and cohort studies comparing surveillance intervals. Pre-defined comparisons included: <3 vs. 4–5 years, <5 vs. 6–10 years, <6 vs. 7–10 years, <10 vs. >10 years. Two reviewers independently performed study selection, data extraction, and quality assessment. Primary outcomes were CRC incidence and CRC mortality. Secondary outcomes included CRC stage, adverse events, all-cause mortality, and advanced adenoma (AA) incidence. Pooled risk ratios (RR) with 95% confidence intervals (CI) were calculated with a random-effects model. Heterogeneity was assessed by Chi2 (P < 0.15) and I2 tests (>25%). We assessed the certainty of evidence (CoE) with the GRADE approach.

From 3086 citations, 9 cohort studies (8 retrospective, 1 prospective) were included. No RCTs were identified. Analysis revealed a dose-dependent relationship between longer surveillance intervals and higher risk both within and between studies. Surveillance at <3 years compared to 4-5 years showed no significant difference in CRC risk (RR 1.06, 95%CI 0.69-1.63). Surveillance at <5 years compared to 6-10 years showed a significantly lower risk of CRC (RR 0.37, 95%CI 0.25-0.54) and AA (RR 0.60, 95%CI 0.53-0.69). This protective effect strengthened as comparator intervals lengthened, with the most pronounced reduction for <10 years versus >10 years for both CRC (RR 0.19, 95%CI 0.12-0.30) and AA (RR 0.53, 95%CI 0.37-0.75). The CoE for all outcomes was very low due to serious risk of bias and serious indirectness.

For patients with 1-2 NAAs, longer surveillance intervals are associated with a dose-dependent increase in CRC and AA risks. These findings argue against de-escalating to average-risk FIT screening and support continued surveillance with colonoscopy.

CPAC (Canadian Partnership Against Cancer)