Automated Determination of Left Ventricular Function Using Electrocardiogram Data in Patients on Maintenance Hemodialysis

Abstract

Background and objectives: Left ventricular ejection fraction is disrupted in patients on maintenance hemodialysis and can be estimated using deep learning models on electrocardiograms. Smaller sample sizes within this population may be mitigated using transfer learning.

Design, setting, participants, & measurements: We identified patients on hemodialysis with transthoracic echocardiograms within 7 days of electrocardiogram using diagnostic/procedure codes. We developed four models: (1) trained from scratch in patients on hemodialysis, (2) pretrained on a publicly available set of natural images (ImageNet), (3) pretrained on all patients not on hemodialysis, and (4) pretrained on patients not on hemodialysis and fine-tuned on patients on hemodialysis. We assessed the ability of the models to classify left ventricular ejection fraction into clinically relevant categories of ≤40%, 41% to ≤50%, and >50%. We compared performance by area under the receiver operating characteristic curve.

Results: We extracted 705,075 electrocardiogram:echocardiogram pairs for 158,840 patients not on hemodialysis used for development of models 3 and 4 and n=18,626 electrocardiogram:echocardiogram pairs for 2168 patients on hemodialysis for models 1, 2, and 4. The transfer learning model achieved area under the receiver operating characteristic curves of 0.86, 0.63, and 0.83 in predicting left ventricular ejection fraction categories of ≤40% (n=461), 41%-50% (n=398), and >50% (n=1309), respectively. For the same tasks, model 1 achieved area under the receiver operating characteristic curves of 0.74, 0.55, and 0.71, respectively; model 2 achieved area under the receiver operating characteristic curves of 0.71, 0.55, and 0.69, respectively, and model 3 achieved area under the receiver operating characteristic curves of 0.80, 0.51, and 0.77, respectively. We found that predictions of left ventricular ejection fraction by the transfer learning model were associated with mortality in a Cox regression with an adjusted hazard ratio of 1.29 (95% confidence interval, 1.04 to 1.59).

Conclusion: A deep learning model can determine left ventricular ejection fraction for patients on hemodialysis following pretraining on electrocardiograms of patients not on hemodialysis. Predictions of low ejection fraction from this model were associated with mortality over a 5-year follow-up period.

Podcast: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2022_06_06_CJN16481221.mp3.

Keywords: electrocardiogram; left ventricular function; maintenance hemodialysis.

Graphical abstract

Figure 1.

Transfer learning overview. ECG, electrocardiogram; HD, hemodialysis.

Figure 2.

Left ventricular ejection fraction (LVEF) prediction in patients on HD. Tested on patients on HD indicates that the model was trained on all non-HD ECG:echocardiogram pairs and tested on patients on HD. Pretrained on ECG data indicate that the ECG transfer learning model was trained on all non-HD ECG:echocardiogram pairs and fine-tuned on patients on HD. (Upper panels) Receiver operating characteristic (ROCs) curves. (Lower panels) Precision-recall (PR) curves. P values were derived from a DeLong test for ROC curves generated from the model trained on patients not on HD and tested on patients on HD and the ECG transfer learning model; they are as follows: LVEF ≤40%, P<0.001; 40% <LVEF ≤50%, P=0.02; LVEF >50%, P<0.001.

Figure 3.

Kaplan–Meier curve of in-hospital mortality in patients on HD with predicted LVEF ≤40%. True positives, false positives, false negatives, and true negatives were ascertained on the basis of the threshold derived from the Youden J Index. P=0.03. True positive denotes patients who were identified by the algorithm as having an LVEF ≤40% and had an LVEF ≤40%. False positive denotes patients who were identified by the algorithm as having an LVEF ≤40% and did not have an LVEF ≤40%. True negative denotes patients who were identified by the algorithm as having an LVEF >40% and had an LVEF >40%. False negative denotes patients who were identified by the algorithm as having an LVEF >40% and had an LVEF ≤40%.