Gerhard-Paul Diller1,2,3, Sonya Babu-Narayan1, Wei Li1, Jelena Radojevic1,4, Aleksander Kempny1, Anselm Uebing1,3,5, Konstantinos Dimopoulos1, Helmut Baumgartner2,3, Michael A Gatzoulis1, Stefan Orwat2,3. 1. Department of Adult Congenital Heart Disease, National Heart and Lung Institute, Imperial College London, Royal Brompton Hospital, Sydney Street, London, UK. 2. Department of Cardiology III, Adult Congenital and Valvular Heart Disease, University Hospital Muenster, Albert-Schweitzer Campus 1, Muenster, Germany. 3. Competence Network for Congenital Heart Defects, DZHK (German Centre for Cardiovascular Research), Berlin, Germany. 4. Cardiologie Congenitale, Strasbourg, France. 5. Division of Paediatric Cardiology, University Hospital Muenster, Albert-Schweitzer Campus 1, Muenster, Germany.
Abstract
AIMS: To investigate the utility of novel deep learning (DL) algorithms in recognizing transposition of the great arteries (TGA) after atrial switch procedure or congenitally corrected TGA (ccTGA) based on routine transthoracic echocardiograms. In addition, the ability of DL algorithms for delineation and segmentation of the systemic ventricle was evaluated. METHODS AND RESULTS: In total, 132 patients (92 TGA and atrial switch and 40 with ccTGA; 60% male, age 38.3 ± 12.1 years) and 67 normal controls (57% male, age 48.5 ± 17.9 years) with routine transthoracic examinations were included. Convolutional neural networks were trained to classify patients by underlying diagnosis and a U-Net design was used to automatically segment the systemic ventricle. Convolutional networks were build based on over 100 000 frames of an apical four-chamber or parasternal short-axis view to detect underlying diagnoses. The DL algorithm had an overall accuracy of 98.0% in detecting the correct diagnosis. The U-Net architecture model correctly identified the systemic ventricle in all individuals and achieved a high performance in segmenting the systemic right or left ventricle (Dice metric between 0.79 and 0.88 depending on diagnosis) when compared with human experts. CONCLUSION: Our study demonstrates the potential of machine learning algorithms, trained on routine echocardiographic datasets to detect underlying diagnosis in complex congenital heart disease. Automated delineation of the ventricular area was also feasible. These methods may in future allow for the longitudinal, objective, and automated assessment of ventricular function. Published on behalf of the European Society of Cardiology. All rights reserved.
AIMS: To investigate the utility of novel deep learning (DL) algorithms in recognizing transposition of the great arteries (TGA) after atrial switch procedure or congenitally corrected TGA (ccTGA) based on routine transthoracic echocardiograms. In addition, the ability of DL algorithms for delineation and segmentation of the systemic ventricle was evaluated. METHODS AND RESULTS: In total, 132 patients (92 TGA and atrial switch and 40 with ccTGA; 60% male, age 38.3 ± 12.1 years) and 67 normal controls (57% male, age 48.5 ± 17.9 years) with routine transthoracic examinations were included. Convolutional neural networks were trained to classify patients by underlying diagnosis and a U-Net design was used to automatically segment the systemic ventricle. Convolutional networks were build based on over 100 000 frames of an apical four-chamber or parasternal short-axis view to detect underlying diagnoses. The DL algorithm had an overall accuracy of 98.0% in detecting the correct diagnosis. The U-Net architecture model correctly identified the systemic ventricle in all individuals and achieved a high performance in segmenting the systemic right or left ventricle (Dice metric between 0.79 and 0.88 depending on diagnosis) when compared with human experts. CONCLUSION: Our study demonstrates the potential of machine learning algorithms, trained on routine echocardiographic datasets to detect underlying diagnosis in complex congenital heart disease. Automated delineation of the ventricular area was also feasible. These methods may in future allow for the longitudinal, objective, and automated assessment of ventricular function. Published on behalf of the European Society of Cardiology. All rights reserved.
Authors: J W Roos-Hesselink; F J Meijboom; S E C Spitaels; R van Domburg; E H M van Rijen; E M W J Utens; J McGhie; E Bos; A J J C Bogers; M L Simoons Journal: Eur Heart J Date: 2004-07 Impact factor: 29.983
Authors: Roberto M Lang; Michelle Bierig; Richard B Devereux; Frank A Flachskampf; Elyse Foster; Patricia A Pellikka; Michael H Picard; Mary J Roman; James Seward; Jack S Shanewise; Scott D Solomon; Kirk T Spencer; Martin St John Sutton; William J Stewart Journal: J Am Soc Echocardiogr Date: 2005-12 Impact factor: 5.251
Authors: Eirik Pettersen; Thomas Helle-Valle; Thor Edvardsen; Harald Lindberg; Hans-Jørgen Smith; Bjarne Smevik; Otto A Smiseth; Kai Andersen Journal: J Am Coll Cardiol Date: 2007-06-11 Impact factor: 24.094
Authors: Andreas P Kalogeropoulos; Vasiliki V Georgiopoulou; Grigorios Giamouzis; Maria-Alexandra Pernetz; Athanasios Anadiotis; Michael McConnell; Stamatios Lerakis; Javed Butler; Wendy M Book; Randolph P Martin Journal: Hellenic J Cardiol Date: 2009 Jul-Aug
Authors: Evangelos P Matsakas; Anastasia S Perpinia; Efterpi H Kambitsi; Haralambos I Kossyvakis; Eftychia S Hamodraka Journal: Hellenic J Cardiol Date: 2005 Sep-Oct
Authors: Gerhard-Paul Diller; Astrid E Lammers; Sonya Babu-Narayan; Wei Li; Robert M Radke; Helmut Baumgartner; Michael A Gatzoulis; Stefan Orwat Journal: Int J Cardiovasc Imaging Date: 2019-07-19 Impact factor: 2.357
Authors: Felix K Wegner; Maria L Benesch Vidal; Philipp Niehues; Kevin Willy; Robert M Radke; Philipp D Garthe; Lars Eckardt; Helmut Baumgartner; Gerhard-Paul Diller; Stefan Orwat Journal: J Clin Med Date: 2022-01-28 Impact factor: 4.241