Validation of the Minnesota Pectoralis Risk Score to predict mortality in the HeartMate 3 population

Background: The Minnesota Pectoralis Risk Score (MPRS) utilizes computed tomography–quantified thoracic muscle and clinical variables to predict survival after left ventricular assist device (LVAD) implantation. The model has not been prospectively tested in HeartMate 3 recipients. Methods: A single-center HeartMate 3 cohort from July 2016 to July 2021 (n = 108) was utilized for this analysis. Cohort subjects with complete covariates for MPRS calculation (pectoralis muscle measures, Black race, creatinine, total bilirubin, body mass index, bridge to transplant status, and presence/absence of contrast) implanted after MPRS development were included. MPRS were calculated on each subject. Receiver operating characteristic curves were generated to test model discrimination at 30-day, 90-day, and 1-year mortality post-LVAD. Next, the performance of the 1-year post-LVAD outcome was compared to the HeartMate 3 survival risk score (HM3RS). Results: The mean age was 58 (15 years), 80% (86/108) were male, and 26% (28/108) were destination therapy. The area under the curve (AUC) for the MPRS model to predict post-LVAD mortality was 0.73 at 30 days, 0.78 at 90 days, and 0.81 at 1 year. The AUC for the HM3RS for the 1-year outcome was 0.693. Each 1-unit point of the MPRS was associated with a significant increase in the hazard rate of death after LVAD (hazard ratio 2.1, 95% confidence interval 1.5-3.0, p < 0.0001). Conclusions: The MPRS had high performance in this prospective validation, particularly with respect to 90-day and 1-year post-LVAD mortality. Such a tool can provide additional information regarding risk stratification to aid informed decision-making.