Rationale Myocardial delivery of human mesenchymal stem cells (hMSCs) is an emerging therapy for treating the failing heart. However, the relative effects of hMSC-mediated heterocellular coupling (HC) and paracrine signaling (PS) on human cardiac contractility and arrhythmogenicity remain unresolved. Objective To better understand hMSC PS and HC effects on human cardiac contractility and arrhythmogenicity by integrating experimental and computational approaches. Methods and Results Extending our previous hMSC-cardiomyocyte HC computational model, we incorporated experimentally calibrated hMSC PS effects on cardiomyocyte L-type calcium channel/SERCA activity and cardiac tissue fibrosis. Excitation-contraction simulations of hMSC PS-only and combined HC+PS effects on human cardiomyocytes were representative of human engineered cardiac tissue (hECT) contractile function measurements under matched experimental treatments. Model simulations and hECTs both demonstrated hMSC-mediated effects were most pronounced under PS-only conditions, where developed force increased approximately 4-fold compared to non-hMSC-supplemented controls during physiologic 1-Hz pacing. Simulations predicted contractility of isolated healthy and ischemic adult human cardiomyocytes would be minimally sensitive to hMSC HC, driven primarily by PS. Dominance of hMSC PS was also revealed in simulations of fibrotic cardiac tissue, where hMSC PS protected from potential pro-arrhythmic effects of HC at various levels of engraftment. Finally, to study the nature of the hMSC paracrine effects on contractility, proteomic analysis of hECT/hMSC conditioned media predicted activation of PI3K/Akt signaling, a recognized target of both soluble and exosomal fractions of the hMSC secretome. Treating hECTs with exosomes-enriched, but not exosomes-depleted, fractions of the hMSC secretome recapitulated the effects observed with hMSC conditioned media on hECT developed force and expression of calcium handling genes (e.g., SERCA2a, L-type calcium channel). Conclusions Collectively, this integrated experimental and computational study helps unravel relative hMSC PS and HC effects on human cardiac contractility and arrhythmogenicity, and provides novel insight into the role of exosomes in hMSC paracrine-mediated effects on contractility.

Study Design. A cross-sectional database study. Objective. The aim of this study was to train and validate machine learning models to identify risk factors for complications following posterior lumbar spine fusion. Summary of Background Data. Machine learning models such as artificial neural networks (ANNs) are valuable tools for analyzing and interpreting large and complex datasets. ANNs have yet to be used for risk factor analysis in orthopedic surgery. Methods. The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for patients who underwent posterior lumbar spine fusion. This query returned 22,629 patients, 70% of whom were used to train our models, and 30% were used to evaluate the models. The predictive variables used included sex, age, ethnicity, diabetes, smoking, steroid use, coagulopathy, functional status, American Society for Anesthesiology (ASA) class ≥3, body mass index (BMI), pulmonary comorbidities, and cardiac comorbidities. The models were used to predict cardiac complications, wound complications, venous thromboembolism (VTE), and mortality. Using ASA class as a benchmark for prediction, area under receiver operating curves (AUC) was used to determine the accuracy of our machine learning models. Results. On the basis of AUC values, ANN and LR both outperformed ASA class for predicting all four types of complications. ANN was the most accurate for predicting cardiac complications, and LR was most accurate for predicting wound complications, VTE, and mortality, though ANN and LR had comparable AUC values for predicting all types of complications. ANN had greater sensitivity than LR for detecting wound complications and mortality. Conclusion. Machine learning in the form of logistic regression and ANNs were more accurate than benchmark ASA scores for identifying risk factors of developing complications following posterior lumbar spine fusion, suggesting they are potentially great tools for risk factor analysis in spine surgery.

This study trained long short-term memory (LSTM) recurrent neural networks (RNNs) incorporating an attention mechanism to predict daily sepsis, myocardial infarction (MI), and vancomycin antibiotic administration over two week patient ICU courses in the MIMIC-III dataset. These models achieved next-day predictive AUC of 0.876 for sepsis, 0.823 for MI, and 0.833 for vancomycin administration. Attention maps built from these models highlighted those times when input variables most influenced predictions and could provide a degree of interpretability to clinicians. These models appeared to attend to variables that were proxies for clinician decision-making, demonstrating a challenge of using flexible deep learning approaches trained with EHR data to build clinical decision support. While continued development and refinement is needed, we believe that such models could one day prove useful in reducing information overload for ICU physicians by providing needed clinical decision support for a variety of clinically important tasks.