Lesion Detection Rates at Diagnostic Colonoscopy: Implications for the New Definition of Adenoma Detection Rate

Background: The extended adenoma detection rate (ADR) definition now includes diagnostic colonoscopies. We hypothesized that ADR and other lesion detection rates are higher for gastrointestinal bleeding/anemia and similar for pain/bowel symptoms vs. first-time screening, and higher for diagnostic colonoscopies without vs. with prior colorectal testing. Methods: We analyzed 36,242 colonoscopies by 60 endoscopists in the Stanford Colonoscopy Quality Assurance Program in patients aged ≥45 years from 2017-2024. ADR, sessile serrated lesion detection rate (SSLDR), and advanced lesion detection rates were compared across colonoscopy indications, accounting for prior colorectal testing. Relative rates (RRs) were estimated with and without adjustment for age and sex. Results: ADR was significantly lower for diagnostic (23.6% [95% CI 22.5–24.7%] overall; nearly identical for bleeding/anemia and pain/bowel symptoms) vs. first-time screening colonoscopies (39.0% [95% CI 38.0–39.9%]; unadjusted RR 0.61 [95% CI 0.57–0.64]), despite older patient age and longer procedure times at diagnostic colonoscopies. Overall extended ADR was very similar to first-screening ADR (37.7% [95% CI 37.2–38.2%]; RR 0.97 [95% CI 0.94-0.99]). Prior colorectal testing (46%) did not meaningfully impact lesion detection rates at diagnostic colonoscopy. Adjusting for age and sex further decreased ADR at diagnostic relative to first-screening colonoscopies. Conclusion: ADR and other lesion detection rates were substantially lower in diagnostic vs. first-screening colonoscopies, irrespective of prior colorectal testing. If symptomatic patients truly harbor below-average colorectal neoplasia risk, then patient mix could affect expanded ADR calculations. Alternatively, if differential inspection quality explains our results, we must ensure that future diagnostic colonoscopies deliver high quality screening.