Detection of left bundle branch block and obstructive coronary artery disease from myocardial perfusion scintigraphy using deep neural networks
(2021) Medical Imaging 2021: Computer-Aided Diagnosis In Progress in Biomedical Optics and Imaging - Proceedings of SPIE 11597.- Abstract
Myocardial perfusion scintigraphy, which is a non-invasive imaging technique, is one of the most common cardiological examinations performed today, and is used for diagnosis of coronary artery disease. Currently the analysis is performed visually by physicians, but this is both a very time consuming and a subjective approach. These are two of the motivations for why an automatic tool to support the decisions would be useful. We have developed a deep neural network which predicts the occurrence of obstructive coronary artery disease in each of the three major arteries as well as left bundle branch block. Since multiple, or none, of these could have a defect, this is treated as a multi-label classification problem. Due to the highly... (More)
Myocardial perfusion scintigraphy, which is a non-invasive imaging technique, is one of the most common cardiological examinations performed today, and is used for diagnosis of coronary artery disease. Currently the analysis is performed visually by physicians, but this is both a very time consuming and a subjective approach. These are two of the motivations for why an automatic tool to support the decisions would be useful. We have developed a deep neural network which predicts the occurrence of obstructive coronary artery disease in each of the three major arteries as well as left bundle branch block. Since multiple, or none, of these could have a defect, this is treated as a multi-label classification problem. Due to the highly imbalanced labels, the training loss is weighted accordingly. The prediction is based on two polar maps, captured during stress in upright and supine position, together with additional information such as BMI and angina symptoms. The polar maps are constructed from myocardial perfusion scintigraphy examinations conducted in a dedicated Cadmium-Zinc-Telluride cardio camera (D-SPECT Spectrum Dynamics). The study includes data from 759 patients. Using 5-fold cross-validation we achieve an area under the receiver operating characteristics curve of 0.89 as average on per-vessel level for the three major arteries, 0.94 on per-patient level and 0.82 for left bundle branch block.
(Less)
- author
- Arvidsson, Ida LU ; Overgaard, Niels Christian LU ; Davidsson, Anette ; Frias Rose, Jeronimo ; Aström, Kalle LU ; Ochoa Figueroa, Miguel and Heyden, Anders LU
- organization
- publishing date
- 2021
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- Convolutional neural network, Deep learning, Left bundle branch block, Obstructive coronary artery disease
- host publication
- Medical Imaging 2021 : Computer-Aided Diagnosis - Computer-Aided Diagnosis
- series title
- Progress in Biomedical Optics and Imaging - Proceedings of SPIE
- editor
- Mazurowski, Maciej A. and Drukker, Karen
- volume
- 11597
- article number
- 115970N
- publisher
- SPIE
- conference name
- Medical Imaging 2021: Computer-Aided Diagnosis
- conference location
- Virtual, Online, United States
- conference dates
- 2021-02-15 - 2021-02-19
- external identifiers
-
- scopus:85103692692
- ISSN
- 1605-7422
- ISBN
- 9781510640238
- DOI
- 10.1117/12.2580890
- project
- Deep learning based evaluation of coronary artery disease and estimation of quantitative coronary angiography using myocardial perfusion imaging
- language
- English
- LU publication?
- yes
- additional info
- Funding Information: Grant support was obtained by Analytic Imaging Diagnostics Arena, Vinnova grant 2017-02447.
- id
- 2bf2460e-8764-473f-8fca-ab9cb23f99fd
- date added to LUP
- 2021-11-29 08:06:36
- date last changed
- 2023-12-22 07:29:55
@inproceedings{2bf2460e-8764-473f-8fca-ab9cb23f99fd, abstract = {{<p>Myocardial perfusion scintigraphy, which is a non-invasive imaging technique, is one of the most common cardiological examinations performed today, and is used for diagnosis of coronary artery disease. Currently the analysis is performed visually by physicians, but this is both a very time consuming and a subjective approach. These are two of the motivations for why an automatic tool to support the decisions would be useful. We have developed a deep neural network which predicts the occurrence of obstructive coronary artery disease in each of the three major arteries as well as left bundle branch block. Since multiple, or none, of these could have a defect, this is treated as a multi-label classification problem. Due to the highly imbalanced labels, the training loss is weighted accordingly. The prediction is based on two polar maps, captured during stress in upright and supine position, together with additional information such as BMI and angina symptoms. The polar maps are constructed from myocardial perfusion scintigraphy examinations conducted in a dedicated Cadmium-Zinc-Telluride cardio camera (D-SPECT Spectrum Dynamics). The study includes data from 759 patients. Using 5-fold cross-validation we achieve an area under the receiver operating characteristics curve of 0.89 as average on per-vessel level for the three major arteries, 0.94 on per-patient level and 0.82 for left bundle branch block.</p>}}, author = {{Arvidsson, Ida and Overgaard, Niels Christian and Davidsson, Anette and Frias Rose, Jeronimo and Aström, Kalle and Ochoa Figueroa, Miguel and Heyden, Anders}}, booktitle = {{Medical Imaging 2021 : Computer-Aided Diagnosis}}, editor = {{Mazurowski, Maciej A. and Drukker, Karen}}, isbn = {{9781510640238}}, issn = {{1605-7422}}, keywords = {{Convolutional neural network; Deep learning; Left bundle branch block; Obstructive coronary artery disease}}, language = {{eng}}, publisher = {{SPIE}}, series = {{Progress in Biomedical Optics and Imaging - Proceedings of SPIE}}, title = {{Detection of left bundle branch block and obstructive coronary artery disease from myocardial perfusion scintigraphy using deep neural networks}}, url = {{http://dx.doi.org/10.1117/12.2580890}}, doi = {{10.1117/12.2580890}}, volume = {{11597}}, year = {{2021}}, }