Nonadvanced Adenoma Number and Size as Predictors of Metachronous Advanced Colorectal Neoplasm: Implications for...

Surveillance for nonadvanced adenomas (NAAs) relies on lesion count, but evidence is limited for individuals with 3 to 4 or 5 to 10 NAAs and for size-based risk stratification. We evaluated advanced colorectal neoplasm (ACN; adenomas ≥10 mm, tubulovillous/villous histology, high-grade dysplasia, or cancer) risk by NAA count and size.

A total of 16,870 adults aged 50 to 75 years with ≥2 colonoscopies (2004-2023) were classified into 4 groups by NAA count and by size (diminutive ≤5 mm vs small 6-9 mm). Over a median 5.3-year follow-up, ACN risk was assessed using inverse probability of censoring weighting and Cox regression.

Participants were categorized by NAA count: 0 (n = 10,023), 1 to 2 (n = 5869), 3 to 4 (n = 781), and 5 to 10 (n = 197). The 5-year cumulative incidences of ACN were 1.3% (95% confidence interval [CI], 1.0%-1.5%), 4.0% (95% CI, 3.4%-4.7%), 5.3% (95% CI, 3.6%-7.8%), and 10.1% (95% CI, 5.8%-17.1%). Compared with no adenomas, hazard ratios were 1.64 (95% CI, 1.36-1.97), 2.75 (95% CI, 1.91-3.97), and 5.21 (95% CI, 2.89-9.38) for 1 to 2, 3 to 4, and 5 to 10 NAAs, respectively. In size-stratified analyses, 5-year ACN incidences were higher with ≥1 small adenoma than diminutive-only within the same count strata: 5.5% (95% CI, 4.0%-7.5%) vs 3.7% (95% CI, 3.0%-4.4%) for 1 to 2 NAAs, and 8.1% (95% CI, 5.1%-12.8%) vs 3.0% (95% CI, 1.5%-5.9%) for 3 to 4 NAAs.

ACN risk increased with NAA count, with the highest risk observed in 5 to 10 NAAs, whereas 3 to 4 NAAs had a higher risk than 1 to 2 NAAs. Small adenomas further elevated risk within count-based groups. NAA number and size should both be considered when determining surveillance intervals.