Simple proxies for detection of clinically significant serrated polyps and data for their benchmarks

Serrated polyp detection rate (SDR) is a potential quality indicator for preventing colorectal cancer associated with the serrated pathway. Using clinically significant SDR (CSSDR) has been suggested based on clinically significant serrated polyp's ability to be colorectal cancer precursors. Correlations between CSSDR and simpler SDRs, other than proximal SDR, have not yet been studied. We aimed to investigate which simpler SDR indicator is most relevant to CSSDR or adenoma detection rate (ADR) and provide benchmark data.

We analyzed 26 627 colonoscopies performed by 30 endoscopists. Clinically significant serrated polyps were defined as any sessile serrated adenoma/polyp or traditional serrated adenoma, hyperplastic polyps ≥ 5 mm in the proximal colon, or hyperplastic polyps ≥ 10 mm anywhere in the colon. Correlation of CSSDR and ADR with other simple SDRs, SDR-pathology (sessile serrated adenoma/polyp or traditional serrated adenoma), SDR-size (≥ 10 mm), and SDR-location (proximal location) was analyzed using Pearson's correlation test and Steiger's z-test.

The CSSDR was 1.7% to 13.2% (mean = 6.1%). The correlation coefficient of CSSDR/SDR-size was 0.91 (P < 0.01), which was higher than that of CSSDR/SDR-location (0.64, P < 0.01) (0.91 vs 0.61, P < 0.01). The correlation coefficient of ADR/CSSDR and ADR/SDR-location was 0.41 (P < 0.01) and 0.81 (P < 0.01), respectively. For ADR ≥ 25%, endoscopists' median screening CSSDR was 5.4%, while SDR-location and SDR-size were 10.9% and 2.2%, respectively.

Large SDR could be a simple proxy for CSSDR, in addition to proximal SDR. Large SDR and proximal SDR benchmarks of 2.2% and 10.9% may guide adequate serrated polyp detection with uniform definitions and simpler calculations.