Micro-ultrasound experience and clinically significant prostate cancer detection: a center-level analysis of the optimum randomized trial

Micro-ultrasound real-time imaging technology is increasingly performed for prostate biopsy. While a learning curve is expected, the effect of operator experience on clinically significant prostate cancer (csPCa) detection has not been clearly defined.

We analyzed data from the multicenter randomized OPTIMUM trial utilizing two logistic regression models to predict csPCa. Center-level experience was represented by two standardized measures derived from operator-level data: Center Case Load Score and Center Mastery Score. Model 1 included the Center Mastery Score and Center Case Volume; Model 2 excluded them. Performance was assessed using confusion matrices and the area under the curve (AUC).

Experience proxies did not improve discrimination for csPCa detection. Sensitivity/specificity/AUC were 0.851/0.523/0.803 vs 0.851/0.544/0.803 for Models 1 and 2, respectively. In Model 1, experience proxies were not significant predictors; csPCa was significantly associated with increasing age (p < 0.001), higher PSA (p < 0.001), smaller prostate volume (p < 0.001), suspicious DRE (p = 0.001). Low PRI-MUS Score (1-2) was associated with decreased risk of csPCa (p = 0.001). Operator heterogeneity and potential bias in experience proxies remain limitations.

Ensuring Level 3 proficiency among all participating operators and incorporating experience proxies with clinical risk factors and PRI-MUS score, center-level experience was not associated with csPCa detection. Our results reflect the successful integration of proficiency standards in clinical practice, while minimizing learning curve effects, and suggest that the required proficiency threshold was an effective proxy for mastery of microUS.