Machine learning algorithms can predict myocardial recovery in patients with LVAD support

Machine learning-based models have greater discriminatory ability compared to conventional regression-based models in predicting the likelihood of myocardial recovery in patients receiving a left ventricular assist device (LVAD), according to the results of a study published in Circulation: heart failure.

Researchers used data from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) to recruit patients ages 18 and older with heart failure who received sustained continuous flow mechanical circulatory support from 2008 to 2017. A total of 20,270 patients were included. in the study. The primary endpoint was explantation of the LVAD for the indication of myocardial salvage.

The researchers extracted 98 raw clinical variables from the INTERMACS dataset to include for feature selection. Machine learning models were developed in the training cohort (70%) and were evaluated in the validation cohort (30%).

Continue reading

Of the 98 variables, 28 with non-zero coefficients were selected for training the machine learning model. Of these, 14 characteristics had a positive association with LVAD-induced myocardial recovery, including bridge-to-recovery implantation strategy, current smoking, postpartum cardiomyopathy, and recent cardiac diagnosis. (1 month to 1 year). In addition, 14 characteristics had a negative association with LVAD-induced myocardial recovery, including use of an implantable cardioverter-defibrillator, post-implantation left ventricular ejection fraction (0% to 20%), and implantation of a right ventricular assist device with LVAD.

The researchers developed 5 machine learning models, including Bayesian Logistic Regression (B-LR), Linear Support Vector Machine, Gradient Boosted Decision Tree, Neural Network, and Random Forest. All of these models showed the ability to predict LVAD-induced myocardial recovery in the validation cohort with an area under the curve (AUC) greater than 0.810.

The discriminatory ability of machine learning models was significantly better than that of regression-based INTERMACS recovery scores, including INTERMACS Cardiac Recovery Score (I-CARS) and INTERMACS Recovery Score (I-TOPS), which had AUCs less than 0.750 (all P <.001>

The researchers also performed additional multivariate logistic regression analysis in the training dataset and developed a new INTERMACS LVAD Recovery Risk Score, as the I-CARS and I-TOPS scores had been derived from earlier versions of the INTERMACS data. The discriminatory ability of the new INTERMACS LVAD recovery risk score (AUC, 0.796) was higher than that of I-CARS and I-TOPS, but it was lower than the best performing B-LR machine learning model (AUC, 0.824 ) in the validation dataset (P =.046).

The cumulative incidence of LVAD explantation for myocardial recovery was significantly increased in patients whose recovery was predicted in machine learning models compared to those who were not (5.1%, 11.5%, 15.8%, and 18.8% versus 0.2%, 1.4%, 1.9%, and 2.6% at 1, 2, 3, and 4 years of LVAD support, respectively; log rank P <.001>

As limitations of the study, the researchers noted that their analysis was limited to clinical variables from the INTERMACS registry and that echocardiographic or radiographic imaging data were not available. Additionally, the machine learning models were validated internally but not externally due to the small number of recovering patients at a given center. Additionally, newer devices such as the Heartmate 3 LVAD were not well represented in the cohort.

“Machine learning tools can help the care team better identify patients who are likely to recover with LVAD support so that recovery efforts can be maximized on these individuals,” the researchers noted.

Disclosure: Some of the study authors have disclosed affiliations with biotechnology, pharmaceutical and/or device companies. Please see the original citation for a full list of author disclosures.

Reference

Topkara VK, Elias P, Jain R, Sayer G, Burkhoff D, Uriel N. Machine learning-based prediction of myocardial recovery in patients with left ventricular assist device. Cardiac failure Circ. Published online December 24, 2021. doi:10.1161/CIRCHEARTFAILURE.121.008711