Log in to MyACC
Machine Learning to Forecast Patient Prognosis After PCI
Study Questions:
What are the limitations of regression-based percutaneous coronary intervention (PCI) risk models by employing machine learning methods to derive discriminatory models that pre-emptively identify patient populations at risk for mortality and rehospitalization after PCI?
Methods:
The investigators evaluated 11,709 distinct patients who underwent 14,349 PCIs between January 2004 and December 2013, in the Mayo Clinic PCI registry. Fifty-two demographic and clinical parameters known at the time of admission were used to predict in-hospital mortality and 358 additional variables available at discharge were examined to identify patients at risk for congestive heart failure (CHF) readmission. For each event, the authors trained a random forest regression model (i.e., machine learning) to estimate the time-to-event. Eight-fold cross-validation was used to estimate model performance. They used the predicted time-to-event as a score, generated a receiver operating characteristic (ROC) curve, and calculated the area under the curve (AUC). Model performance was then compared to a logistic regression model using pairwise comparisons of AUCs and calculation of net reclassification indices (NRIs).
Results:
The predictive algorithm identified a high-risk cohort representing 2% of all patients who had an in-hospital mortality of 45.5% (95% confidence interval [CI], 43.5-47.5%) compared to a risk of 2.1% for the general population (AUC, 0.925; 95% CI, 0.92-0.93). Advancing age, CHF, and shock on presentation were the leading predictors for the outcome. A high-risk group representing 1% of all patients was identified with 30-day CHF rehospitalization of 8.1% (95% CI, 6.3-10.2%). Random forest regression outperformed logistic regression for predicting 30-day CHF readmission (AUC 0.90 vs. 0.85, p = 0.003; NRI: 5.14%) and 180-day cardiovascular death (AUC 0.88 vs. 0.81, p = 0.02; NRI 0.02%).
Conclusions:
The authors concluded that random forest regression models (machine learning) were more predictive and discriminative than standard regression methods at identifying patients at risk for 180-day cardiovascular mortality and 30-day CHF rehospitalization, but not in-hospital mortality.
Perspective:
This study reports that cross-validated machine learned random forest algorithms were more predictive and discriminative than standard regression methods at identifying patients at the risk for post-procedure mortality and 30-day CHF rehospitalization. Furthermore, machine learning was also effective at identifying patient subgroups at high risk for post-procedure complications and readmission. While, the current study supports a potential role for the integration of machine learning into clinical practice for the identification of high-risk populations following PCI, given that machine learning results were validated internally in this study, external validation is needed.
Clinical Topics: Acute Coronary Syndromes, Heart Failure and Cardiomyopathies, Invasive Cardiovascular Angiography and Intervention, Prevention, Acute Heart Failure, Interventions and ACS
Keywords: Acute Coronary Syndrome, Area Under Curve, Heart Failure, Hospital Mortality, Learning, Patient Discharge, Patient Readmission, Percutaneous Coronary Intervention, Primary Prevention, Risk Factors, ROC Curve, Shock, Cardiogenic
< Back to Listings
