Prostate MRI learning curves: establishing training benchmarks for radiology and urology trainees.

OBJECTIVES

Evidence-based training benchmarks for prostate multiparametric MRI (mpMRI) interpretation remain undefined amid growing educational demands. We compared learning curves between radiology and urology trainees and quantified the impact of prior radiological experience.

MATERIALS AND

METHODS

Fourteen trainees (10 radiology, median 2.7 years experience; 4 urology, no imaging experience), all naïve to prostate mpMRI, prospectively interpreted 200 cases using a feedback-based platform. Performance metrics included agreement with expert consensus reference for PI-RADSv2.1 (≥ 3), PI-QUALv2 image quality, extraprostatic extension (EPE) grading, and readout time. Learning curves were modeled using generalized estimating equations; segmented regression identified inflection points; bootstrapping generated 95% CIs.

RESULTS

Prior radiological experience showed no significant impact on PI-RADSv2.1 (OR per year 1.06 [95% CI: 0.96, 1.16]) or PI-QUALv2 (1.05 [0.99, 1.23]), with a minor effect on EPE grading (1.11 [1.03, 1.24]). Final PI-RADSv2.1 agreement with reference was similar (urology 80.9%, radiology 77.4%; OR 1.24 [0.55, 3.10]), with sensitivity/specificity 0.84/0.80 and 0.83/0.79, and Cohen's κ values (0.64 and 0.61) matching inter-expert κ = 0.63. Learning plateaued after 69-75 cases. Urology trainees demonstrated higher baseline PI-QUALv2/EPE agreement (OR 2.01 [1.35, 3.02] and 1.90 [1.11, 2.93]), but radiology trainees achieved similar final performance (PI-QUALv2: 88.0% vs 89.9%, OR 0.82 [0.35, 1.72]; EPE: 84.6% vs 90.0%, OR 0.61 [0.31, 1.42]). Readout times decreased markedly in both groups (final difference 53.3 s [-9.4, 95.9]).

CONCLUSION

Feedback-based training enabled similar prostate mpMRI interpretation performance across specialties, with most learning within 75 cases. Prior radiological experience had a limited impact. These empirical benchmarks inform certification standards and early-residency curricula in radiology and urology.

KEY POINTS: Question Evidence-based training benchmarks for prostate mpMRI interpretation competency in radiology and urology trainees remain undefined, despite growing educational needs and clinical demands. Findings Learning curves of 200 prostate mpMRI cases with feedback showed radiology and urology trainees plateauing after 69-75 cases with similar PI-RADSv2.1, PI-QUALv2, EPE grading performance. Clinical relevance Our findings establish an empirical benchmark (~75 cases) to guide prostate mpMRI certification standards and support the implementation of training curricula early in residency across specialties, regardless of prior radiological experience.