StrainNet: Improved Myocardial Strain Analysis of Cine MRI by Deep Learning from DENSE

Yu Wang; Changyu Sun; Sona Ghadimi; Daniel C Auger; Pierre Croisille; Magalie Viallon; Kenneth Mangion; Colin Berry; Christopher M Haggerty; Linyuan Jing; Brandon K Fornwalt; J Jane Cao; Joshua Cheng; Andrew D Scott; Pedro F Ferreira; John N Oshinski; Daniel B Ennis; Kenneth C Bilchick; Frederick H Epstein

doi:10.1148/ryct.220196

StrainNet: Improved Myocardial Strain Analysis of Cine MRI by Deep Learning from DENSE

Radiol Cardiothorac Imaging. 2023 May 4;5(3):e220196. doi: 10.1148/ryct.220196. eCollection 2023 Jun.

Authors

Yu Wang¹, Changyu Sun¹, Sona Ghadimi¹, Daniel C Auger¹, Pierre Croisille¹, Magalie Viallon¹, Kenneth Mangion¹, Colin Berry¹, Christopher M Haggerty¹, Linyuan Jing¹, Brandon K Fornwalt¹, J Jane Cao¹, Joshua Cheng¹, Andrew D Scott¹, Pedro F Ferreira¹, John N Oshinski¹, Daniel B Ennis¹, Kenneth C Bilchick¹, Frederick H Epstein¹

Affiliation

¹ From the Department of Biomedical Engineering, University of Virginia, Biomedical Engineering and Medical Science Building, Room 2013, MR5, Charlottesville, VA 22903 (Y.W., C.S., S.G., D.C.A., F.H.E.); Department of Biomedical, Biological and Chemical Engineering and Department of Radiology, University of Missouri, Columbia, Mo (C.S.); Department of Radiology, University Hospital of Saint Etienne, Saint Etienne, France (P.C.); CREATIS (UMR CNRS 5220, U1206 INSERM), INSA Lyon, Lyon, France (P.C., M.V.); BHF Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, Scotland (K.M., C.B.); Department of Translational Data Science and Informatics, Geisinger Health System, Danville, Pa (C.M.H., L.J., B.K.F.); Cardiovascular Research Center, University of Kentucky, Lexington, Ky (C.M.H., L.J., B.K.F.); The Heart Center, St Francis Hospital, Roslyn, NY (J.J.C., J.C.); Cardiovascular Magnetic Resonance Unit, The Royal Brompton Hospital and National Heart and Lung Institute, Imperial College London, London, England (A.D.S., P.F.F.); Department of Radiology & Imaging Sciences and Biomedical Engineering, Emory University, Atlanta, Ga (J.N.O.); Department of Radiology, Stanford University, Stanford, Calif (D.B.E.); Department of Medicine (K.C.B.) and Department of Radiology and Medical Imaging (F.H.E.), University of Virginia Health System, Charlottesville, Va.

Abstract

Purpose: To develop a three-dimensional (two dimensions + time) convolutional neural network trained with displacement encoding with stimulated echoes (DENSE) data for displacement and strain analysis of cine MRI.

Materials and methods: In this retrospective multicenter study, a deep learning model (StrainNet) was developed to predict intramyocardial displacement from contour motion. Patients with various heart diseases and healthy controls underwent cardiac MRI examinations with DENSE between August 2008 and January 2022. Network training inputs were a time series of myocardial contours from DENSE magnitude images, and ground truth data were DENSE displacement measurements. Model performance was evaluated using pixelwise end-point error (EPE). For testing, StrainNet was applied to contour motion from cine MRI. Global and segmental circumferential strain (E_cc) derived from commercial feature tracking (FT), StrainNet, and DENSE (reference) were compared using intraclass correlation coefficients (ICCs), Pearson correlations, Bland-Altman analyses, paired t tests, and linear mixed-effects models.

Results: The study included 161 patients (110 men; mean age, 61 years ± 14 [SD]), 99 healthy adults (44 men; mean age, 35 years ± 15), and 45 healthy children and adolescents (21 males; mean age, 12 years ± 3). StrainNet showed good agreement with DENSE for intramyocardial displacement, with an average EPE of 0.75 mm ± 0.35. The ICCs between StrainNet and DENSE and FT and DENSE were 0.87 and 0.72, respectively, for global E_cc and 0.75 and 0.48, respectively, for segmental E_cc. Bland-Altman analysis showed that StrainNet had better agreement than FT with DENSE for global and segmental E_cc.

Conclusion: StrainNet outperformed FT for global and segmental E_cc analysis of cine MRI.Keywords: Image Postprocessing, MR Imaging, Cardiac, Heart, Pediatrics, Technical Aspects, Technology Assessment, Strain, Deep Learning, DENSE Supplemental material is available for this article. © RSNA, 2023.

Keywords: Cardiac; DENSE; Deep Learning; Heart; Image Postprocessing; MR Imaging; Pediatrics; Strain; Technical Aspects; Technology Assessment.

Grants and funding

RG/19/1/34160/BHF_/British Heart Foundation/United Kingdom