Segmentation-based cardiomegaly detection based on semi-supervised estimation of cardiothoracic ratio.

Autor: Thiam P; Institute of Medical Systems Biology, Albert-Einstein-Allee 11, 89081, Ulm, Germany., Kloth C; Department of Diagnostic and Interventional Radiology, Ulm University Medical Center, Albert-Einstein-Allee 23, 89081, Ulm, Germany., Blaich D; Department of Diagnostic and Interventional Radiology, Ulm University Medical Center, Albert-Einstein-Allee 23, 89081, Ulm, Germany., Liebold A; Department of Cardiothoraxic and Vascular Surgery, Ulm University Medical Center, Albert-Einstein-Allee 23, 89081, Ulm, Germany., Beer M; Department of Diagnostic and Interventional Radiology, Ulm University Medical Center, Albert-Einstein-Allee 23, 89081, Ulm, Germany., Kestler HA; Institute of Medical Systems Biology, Albert-Einstein-Allee 11, 89081, Ulm, Germany. hans.kestler@uni-ulm.de.
Jazyk: angličtina
Zdroj: Scientific reports [Sci Rep] 2024 Mar 08; Vol. 14 (1), pp. 5695. Date of Electronic Publication: 2024 Mar 08.
DOI: 10.1038/s41598-024-56079-1
Abstrakt: The successful integration of neural networks in a clinical setting is still uncommon despite major successes achieved by artificial intelligence in other domains. This is mainly due to the black box characteristic of most optimized models and the undetermined generalization ability of the trained architectures. The current work tackles both issues in the radiology domain by focusing on developing an effective and interpretable cardiomegaly detection architecture based on segmentation models. The architecture consists of two distinct neural networks performing the segmentation of both cardiac and thoracic areas of a radiograph. The respective segmentation outputs are subsequently used to estimate the cardiothoracic ratio, and the corresponding radiograph is classified as a case of cardiomegaly based on a given threshold. Due to the scarcity of pixel-level labeled chest radiographs, both segmentation models are optimized in a semi-supervised manner. This results in a significant reduction in the costs of manual annotation. The resulting segmentation outputs significantly improve the interpretability of the architecture's final classification results. The generalization ability of the architecture is assessed in a cross-domain setting. The assessment shows the effectiveness of the semi-supervised optimization of the segmentation models and the robustness of the ensuing classification architecture.
(© 2024. The Author(s).)
Databáze: MEDLINE
Externí odkaz: