Akhil Vaid1,2,3,4, Joy J Jiang4, Ashwin Sawant4,5, Karandeep Singh6, Patricia Kovatch1,3, Alexander W Charney1,3, David M Charytan7, Jasmin Divers8, Benjamin S Glicksberg1,2,3,4, Lili Chan4,9,10, Girish N Nadkarni11,2,4,9,10. 1. Mount Sinai Clinical Intelligence Center, Icahn School of Medicine at Mount Sinai, New York, New York. 2. The Hasso Plattner Institute for Digital Health at Mount Sinai, Icahn School of Medicine at Mount Sinai, New York, New York. 3. Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, New York. 4. Division of Data Driven and Digital Medicine, Icahn School of Medicine at Mount Sinai, New York, New York. 5. Division of Hospital Medicine, Icahn School of Medicine at Mount Sinai, New York, New York. 6. Department of Learning Health Systems, University of Michigan Medical School, Ann Arbor, Michigan. 7. Division of Nephrology, Department of Medicine, New York University Langone Medical Center and New York University Grossman School of Medicine, New York, New York. 8. Division of Health Services, Department of Medicine, New York University Langone Medical Center, New York, New York. 9. Division of Nephrology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York. 10. The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, New York. 11. Mount Sinai Clinical Intelligence Center, Icahn School of Medicine at Mount Sinai, New York, New York girish.nadkarni@mountsinai.org.
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.
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.
Authors: Christopher M Haggerty; Brandon K Fornwalt; Sushravya Raghunath; Alvaro E Ulloa Cerna; Linyuan Jing; David P vanMaanen; Joshua Stough; Dustin N Hartzel; Joseph B Leader; H Lester Kirchner; Martin C Stumpe; Ashraf Hafez; Arun Nemani; Tanner Carbonati; Kipp W Johnson; Katelyn Young; Christopher W Good; John M Pfeifer; Aalpen A Patel; Brian P Delisle; Amro Alsaid; Dominik Beer Journal: Nat Med Date: 2020-05-11 Impact factor: 53.440
Authors: Biykem Bozkurt; Andrew Js Coats; Hiroyuki Tsutsui; Magdy Abdelhamid; Stamatis Adamopoulos; Nancy Albert; Stefan D Anker; John Atherton; Michael Böhm; Javed Butler; Mark H Drazner; G Michael Felker; Gerasimos Filippatos; Gregg C Fonarow; Mona Fiuzat; Juan-Esteban Gomez-Mesa; Paul Heidenreich; Teruhiko Imamura; James Januzzi; Ewa A Jankowska; Prateeti Khazanie; Koichiro Kinugawa; Carolyn S P Lam; Yuya Matsue; Marco Metra; Tomohito Ohtani; Massimo Francesco Piepoli; Piotr Ponikowski; Giuseppe M C Rosano; Yasushi Sakata; Petar SeferoviĆ; Randall C Starling; John R Teerlink; Orly Vardeny; Kazuhiro Yamamoto; Clyde Yancy; Jian Zhang; Shelley Zieroth Journal: J Card Fail Date: 2021-03-01 Impact factor: 5.712