Objective The Model for End- Stage Liver Disease (MELD) score is a composite number of physiologic parameters and likely has non-linear effects on operative outcomes. . We use machine learning to evaluate the relationship between MELD score and outcomes of cardiac surgery. Methods All STS indexed elective cardiac surgical procedures at our institution between 2011 and 2018 were included. MELD score was retrospectively calculated. Logistic regression models and an imbalanced random forest classifier was created on operative mortality using 30 preoperative characteristics. Cox regression models and random forest survival models were created for long-term survival. Variable importance analysis (VIMP) was conducted to rank variables by predictive power. Linear and machine learned models were compared with their receiver operating characteristic (ROC) and Brier score respectively. Results The patient population included 3,872 individuals. Operative mortality was 1.7% and 5-year survival was 82.1%. MELD score was the 4th largest positive predictor on VIMP analysis for both operative long-term survival and the strongest negative predictor for operative mortality. The logistic model ROC area was 0.762, compared to the random forest classifier ROC of 0.674. The Brier score of the random forest survival model was larger (worse) than the cox regression starting at 2 years and continuing throughout the study period. Conclusions MELD score and other continuous variables had high degrees of non-linearity to mortality. This is demonstrated by the fact that MELD score was not significant in the cox multivariable regression but was strongly important in the random forest survival model.

Heart Transplantation in the Current Era: Thinking Outside the Box of Normal Donor CriteriaNicholas R. Hess MD1, Arman Kilic MD2

Background: This study evaluated the utilization and outcomes of postcardiotomy mechanical circulatory support (MCS). Methods: This was a retrospective, single institution analysis of adult cardiac surgery cases that required de novo MCS following surgery from 2011-2018. Patients that were bridged with MCS to surgery were excluded. The primary outcomes were early operative mortality and longitudinal survival. Secondary outcomes included postoperative complications, and five-year all-cause readmission. Results: 533 patients required de novo postcardiotomy MCS, with the most commonly performed procedure being isolated coronary artery bypass grafting (29.8%). Median cardiopulmonary bypass and cross clamp times were 185 (IQR 123-260) minutes and 122 (IQR 81-179) minutes, respectively. A total of 442 (82.9%) of patients were supported with intra-aortic balloon pump counterpulsation, 23 (4.3%) with an Impella device, and 115 (21.6%) with extracorporeal membrane oxygenation. Three (0.6%) patients had an unplanned ventricular assist device placed. Operative mortality was 29.8%. Longitudinal survival was 56.1% and 43.0% at 1- and 5-years, respectively. Survival was lowest in those supported with ECMO and highest with those supported with an Impella (P<0.001). Freedom from readmission was 61.4% at 5-years. Postoperative ECMO was an independent predictor of mortality (HR 5.1, 95% CI 2.0-12.9, P<0.001), but none of the MCS types predicted long-term hospital readmission after risk adjustment. Conclusions: Postcardiotomy MCS is associated with high operative mortality. Even patients that survive to discharge have compromised longitudinal survival, with nearly only half surviving to 1-year. Close follow-up and early referral to advanced heart failure specialists may be prudent in improving these outcomes.

Background: This study investigates the use of modern machine learning (ML) techniques to improve prediction of survival after orthotopic heart transplantation (OHT). Methods: Retrospective study of adult patients undergoing primary, isolated OHT between 2000-2019 as identified in the United Network for Organ Sharing (UNOS) registry. The primary outcome was one-year post-transplant survival. Patients were randomly divided into training (80%) and validation (20%) sets. Dimensionality reduction and data re-sampling were employed during training. Multiple machine learning algorithms were combined into a final ensemble ML model. Discriminatory capability was assessed using area under receiver-operating-characteristic curve (AUROC), net reclassification index (NRI), and decision curve analysis (DCA). Results: A total of 33,657 OHT patients were evaluated. One-year mortality was 11% (n=3,738). In the validation cohort, the AUROC of singular logistic regression was 0.649 (95% CI 0.628-0.670) compared to 0.691 (95% CI 0.671-0.711) with random forest, 0.691 (95% CI 0.671-0.712) with deep neural network, and 0.653 (95% CI 0.632-0.674) with Adaboost. A final ensemble ML model was created that demonstrated the greatest improvement in AUROC: 0.764 (95% CI 0.745-0.782) (p<0.001). The ensemble ML model improved predictive performance by 72.9% ±3.8% (p<0.001) as assessed by NRI compared to logistic regression. DCA showed the final ensemble method improved risk prediction across the entire spectrum of predicted risk as compared to all other models (p<0.001). Conclusions: Modern ML techniques can improve risk prediction in OHT compared to traditional approaches. This may have important implications in patient selection, programmatic evaluation, allocation policy, and patient counseling and prognostication.

Background: The introduction of integrated thoracic surgery residency programs has led to increased recruitment efforts of medical students to pursue a career in cardiac surgery. With little representation of cardiac surgery in medical school curriculum, we assessed a cardiac surgery mini-elective’s efficacy in improving perceived knowledge among medical students. Methods: Preclinical medical students were offered the opportunity to participate in a cardiac surgery mini-elective, which consisted of five 2-hour sessions. These sessions consisted of didactic and simulation components and covered topics including cardiopulmonary bypass (CPB) and extracorporeal membrane oxygenation (ECMO), aortic disease, aortic valve replacement (AVR), transplant and left ventricular assist devices (LVAD), and coronary artery bypass grafting (CABG). Students completed pre- and post-session survey’s describing their perceived knowledge in these topics. Results: Overall, 22 students completed at least one session of the mini-elective. Fourteen (73.7%) of the students were male. Fifteen (68.2%) students completed at least three out of five sessions. The post-session survey responses showed significantly higher perceived knowledge compared to pre-session responses for all survey prompts of all five sessions. The CPB/ECMO and aortic disease sessions showed the greatest increase in post-session familiarity and perceived knowledge after the session (p<0.001) compared to the CABG, AVR, and transplant/LVAD sessions (p<0.05). Conclusions: Beyond developing interest in cardiac surgery, these data indicate that a well-planned didactic and surgical simulation program may build confidence in students’ knowledge of various cardiac surgical topics. Further studies will need to address how this increase in perceived ability lasts over time and impacts career selection.

Response to Letter to the Editor: “Impact of preoperative troponin levels on cardiac function following coronary surgery for myocardial infarction”Nicholas Hess MD1, Arman Kilic MD2