Background Although second‐generation Murray law–based quantitative flow ratio uses artificial intelligence–driven automation to improve angiography‐derived physiological assessment, the impact of operator learning curves on its diagnostic reliability in routine practice remains unclear. Methods Consecutive patients with suspected myocardial ischemia (May 2021 to July 2024) underwent online Murray law–based quantitative flow ratio computation from angiographic images, followed by fractional flow reserve measurement. Diagnostic accuracy was evaluated using receiver operating characteristic analysis. The number of vessels needed to achieve high performance (area under the curve AUC ≥0.90) was determined. Learning curves were assessed by sequentially expanding the data set starting from the first 11 vessels. First‐order AUC differences (ΔAUC n+1 =AUC n+1 −AUC n ) were calculated to assess the stabilization of diagnostic performance, where AUC n represents the area under the curve after analyzing the first n vessels. Results Murray law–based quantitative flow ratio showed excellent diagnostic accuracy: Vessel‐level AUC was 0.92 (95% CI, 0.89–0.94), and patient‐level AUC was 0.91 (95% CI, 0.88–0.94). An AUC ≥0.90 was achieved after analyzing 170 vessels, with continued improvement thereafter. Peak performance and stability were reached at 441 vessels, where incremental ΔAUC values showed no significant deviation from 0 (Mann–Whitney U test, P >0.05). Conclusions Operator experience significantly affects Murray law–based quantitative flow ratio accuracy. Analyzing at least 170 vessels ensures high reliability, with optimal performance achieved around 441 cases. Structured training is essential for integrating artificial intelligence–enhanced coronary physiology tools into clinical practice.
Deng et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: