A simple scoring system to estimate adenoma detection rate from polyp detection: a two-center observational study

The calculation of adenoma detection rate (ADR) requires pathological information, which makes it time-consuming. We aimed to develop a scoring system to estimate ADR from polyp detection accurately.

The two-center retrospective study included colonoscopy and pathological data, which were divided into training, internal, and external validation cohorts. The factors associated with pathologically confirmed adenoma detection in patients with polyps were identified using the logistic regression model. The scores were assigned based on β-coefficient. The performance of the scoring system was evaluated by the Bland-Altman plot.

Patient age, maximum size of polyps, and number of proximal or distal polyps of any size, except distal diminutive polyps, were significantly associated with pathologically confirmed adenoma detection in patients with polyps. Based on regression coefficients, scores were assigned as follows: age ≥ 70 years, positive; any proximal polyps, positive; distal polyps ≥ 10 mm, positive; age 60 ~ 69 years, 1 point; per distal small polyps, 1 point. Pathologically confirmed adenoma detection was predicted when a patient met any positive criterion or, in the absence of these positive items, had a total score > 1 point. The scoring system showed a high proportion of acceptable estimations when the difference between estimated and actual ADR was less than 2%. In clinically relevant conditions with a small number of samples, the proportion of acceptable estimations rose gradually and achieved its highest level when the number of samples increased to 300.

Our scoring system performed well in estimating ADR from polyp detection. 300 colonoscopies per endoscopist was the best number of samples when the scoring system was used.