Development and validation of a risk scoring system based on baseline angiographic results by visual estimation for risk prEdiction of side-branch OccLusion in coronary bifurcation InterVEntion: The baseline V-RESOLVE score.

Abstract

We aimed to adapt the visual estimation for Risk prEdiction of side-branch OccLusion in coronary bifurcation interVEntion (V-RESOLVE) score to enable risk prediction for side-branch (SB) occlusion using only baseline coronary angiogram data. The V-RESOLVE score, which comprises angiographic factors at baseline and after lesion preparation, is a validated tool for predicting SB occlusion risk in coronary bifurcation intervention. To develop and validate the new scoring system, we used data pertaining to 1,545 patients and another 1,286 consecutive patients. Baseline V-RESOLVE was derived from V-RESOLVE by replacing the two pre-stenting angiographic factors with the corresponding preprocedural characteristics, while maintaining the scoring standard itself. We evaluated the diagnostic performance of baseline V-RESOLVE for predicting SB occlusion and preformed risk stratification with characterization of non-high-risk and high-risk lesions. The area under the receiver operating characteristic curves was similar between baseline V-RESOLVE and V-RESOLVE (0.735 vs 0.756, P = 0.191), with good calibration for baseline V-RESOLVE (Hosmer-Lemeshow P = 0.714). Upon categorization by the baseline V-RESOLVE score, high-risk lesions (score: 14-43) demonstrated significantly higher rate of SB occlusion than did non-high-risk lesions (score: 0-13) (17.31% vs 4.74%, P < 0.01). Considering the V-RESOLVE-based risk stratification as reference, baseline V-RESOLVE had an integrated discrimination index of -1.81% (P = 0.052), and net reclassification improvement of -3.34% (P = 0.509). Upon validation, baseline V-RESOLVE provided satisfactory diagnostic performance and risk stratification. Baseline V-RESOLVE predicts SB occlusion in coronary bifurcation intervention based solely on the preprocedural angiographic results.