Zhuo He1, Xinwei Zhang2, Chen Zhao1, Xing Ling3, Saurabh Malhotra4,5, Zhiyong Qian2, Yao Wang2, Xiaofeng Hou2, Jiangang Zou6, Weihua Zhou7,8. 1. College of Computing, Michigan Technological University, 1400 Townsend Drive, Houghton, MI, USA. 2. Department of Cardiology, The First Affiliated Hospital of Nanjing Medical University, Guangzhou Road 300, Nanjing, 210029, Jiangsu, China. 3. Department of Mathematical Sciences, Michigan Technological University, Houghton, MI, USA. 4. Division of Cardiology, Cook County Health and Hospitals System, Chicago, IL, USA. 5. Division of Cardiology, Rush Medical College, Chicago, IL, USA. 6. Department of Cardiology, The First Affiliated Hospital of Nanjing Medical University, Guangzhou Road 300, Nanjing, 210029, Jiangsu, China. jgzou@njmu.edu.cn. 7. College of Computing, Michigan Technological University, 1400 Townsend Drive, Houghton, MI, USA. whzhou@mtu.edu. 8. Center for Biocomputing and Digital Health, Institute of Computing and Cybersystems, Health Research Institute, Michigan Technological University, Houghton, MI, 49931, USA. whzhou@mtu.edu.
Abstract
BACKGROUND: Studies have shown that the conventional parameters characterizing left ventricular mechanical dyssynchrony (LVMD) measured on gated SPECT myocardial perfusion imaging (MPI) have their own statistical limitations in predicting cardiac resynchronization therapy (CRT) response. The purpose of this study is to discover new predictors from the polarmaps of LVMD by deep learning to help select heart failure patients with a high likelihood of response to CRT. METHODS: One hundred and fifty-seven patients who underwent rest gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at 6 [Formula: see text] 1 month follow up. The autoencoder (AE) technique, an unsupervised deep learning method, was applied to the polarmaps of LVMD to extract new predictors characterizing LVMD. Pearson correlation analysis was used to explain the relationships between new predictors and existing clinical parameters. Patients from the IAEA VISION-CRT trial were used for an external validation. Heatmaps were used to interpret the AE-extracted feature. RESULTS: Complete data were obtained in 130 patients, and 68.5% of them were classified as CRT responders. After variable selection by feature importance ranking and correlation analysis, one AE-extracted LVMD predictor was included in the statistical analysis. This new AE-extracted LVMD predictor showed statistical significance in the univariate (OR 2.00, P = .026) and multivariate (OR 1.11, P = .021) analyses, respectively. Moreover, the new AE-extracted LVMD predictor not only had incremental value over PBW and significant clinical variables, including QRS duration and left ventricular end-systolic volume (AUC 0.74 vs 0.72, LH 7.33, P = .007), but also showed encouraging predictive value in the 165 patients from the IAEA VISION-CRT trial (P < .1). The heatmaps for calculation of the AE-extracted predictor showed higher weights on the anterior, lateral, and inferior myocardial walls, which are recommended as LV pacing sites in clinical practice. CONCLUSIONS: AE techniques have significant value in the discovery of new clinical predictors. The new AE-extracted LVMD predictor extracted from the baseline gated SPECT MPI has the potential to improve the prediction of CRT response.
BACKGROUND: Studies have shown that the conventional parameters characterizing left ventricular mechanical dyssynchrony (LVMD) measured on gated SPECT myocardial perfusion imaging (MPI) have their own statistical limitations in predicting cardiac resynchronization therapy (CRT) response. The purpose of this study is to discover new predictors from the polarmaps of LVMD by deep learning to help select heart failure patients with a high likelihood of response to CRT. METHODS: One hundred and fifty-seven patients who underwent rest gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at 6 [Formula: see text] 1 month follow up. The autoencoder (AE) technique, an unsupervised deep learning method, was applied to the polarmaps of LVMD to extract new predictors characterizing LVMD. Pearson correlation analysis was used to explain the relationships between new predictors and existing clinical parameters. Patients from the IAEA VISION-CRT trial were used for an external validation. Heatmaps were used to interpret the AE-extracted feature. RESULTS: Complete data were obtained in 130 patients, and 68.5% of them were classified as CRT responders. After variable selection by feature importance ranking and correlation analysis, one AE-extracted LVMD predictor was included in the statistical analysis. This new AE-extracted LVMD predictor showed statistical significance in the univariate (OR 2.00, P = .026) and multivariate (OR 1.11, P = .021) analyses, respectively. Moreover, the new AE-extracted LVMD predictor not only had incremental value over PBW and significant clinical variables, including QRS duration and left ventricular end-systolic volume (AUC 0.74 vs 0.72, LH 7.33, P = .007), but also showed encouraging predictive value in the 165 patients from the IAEA VISION-CRT trial (P < .1). The heatmaps for calculation of the AE-extracted predictor showed higher weights on the anterior, lateral, and inferior myocardial walls, which are recommended as LV pacing sites in clinical practice. CONCLUSIONS: AE techniques have significant value in the discovery of new clinical predictors. The new AE-extracted LVMD predictor extracted from the baseline gated SPECT MPI has the potential to improve the prediction of CRT response.
Authors: Maureen M Henneman; Ji Chen; Petra Dibbets-Schneider; Marcel P Stokkel; Gabe B Bleeker; Claudia Ypenburg; Ernst E van der Wall; Martin J Schalij; Ernest V Garcia; Jeroen J Bax Journal: J Nucl Med Date: 2007-06-15 Impact factor: 10.057
Authors: Andrew E Epstein; John P DiMarco; Kenneth A Ellenbogen; N A Mark Estes; Roger A Freedman; Leonard S Gettes; A Marc Gillinov; Gabriel Gregoratos; Stephen C Hammill; David L Hayes; Mark A Hlatky; L Kristin Newby; Richard L Page; Mark H Schoenfeld; Michael J Silka; Lynne Warner Stevenson; Michael O Sweeney; Cynthia M Tracy; Andrew E Epstein; Dawood Darbar; John P DiMarco; Sandra B Dunbar; N A Mark Estes; T Bruce Ferguson; Stephen C Hammill; Pamela E Karasik; Mark S Link; Joseph E Marine; Mark H Schoenfeld; Amit J Shanker; Michael J Silka; Lynne Warner Stevenson; William G Stevenson; Paul D Varosy Journal: J Am Coll Cardiol Date: 2012-12-19 Impact factor: 24.094
Authors: Mati Friehling; Ji Chen; Samir Saba; Raveen Bazaz; David Schwartzman; Evan C Adelstein; Ernest Garcia; William Follansbee; Prem Soman Journal: Circ Cardiovasc Imaging Date: 2011-07-19 Impact factor: 7.792
Authors: Eugene S Chung; Angel R Leon; Luigi Tavazzi; Jing-Ping Sun; Petros Nihoyannopoulos; John Merlino; William T Abraham; Stefano Ghio; Christophe Leclercq; Jeroen J Bax; Cheuk-Man Yu; John Gorcsan; Martin St John Sutton; Johan De Sutter; Jaime Murillo Journal: Circulation Date: 2008-05-05 Impact factor: 29.690
Authors: Serge D Van Kriekinge; Hidetaka Nishina; Muneo Ohba; Daniel S Berman; Guido Germano Journal: J Nucl Med Date: 2008-10-16 Impact factor: 10.057
Authors: Piotr Ponikowski; Adriaan A Voors; Stefan D Anker; Héctor Bueno; John G F Cleland; Andrew J S Coats; Volkmar Falk; José Ramón González-Juanatey; Veli-Pekka Harjola; Ewa A Jankowska; Mariell Jessup; Cecilia Linde; Petros Nihoyannopoulos; John T Parissis; Burkert Pieske; Jillian P Riley; Giuseppe M C Rosano; Luis M Ruilope; Frank Ruschitzka; Frans H Rutten; Peter van der Meer Journal: Eur Heart J Date: 2016-05-20 Impact factor: 29.983