Deep learning-based detection of ascending aortic dilatation on chest radiographs: A diagnostic study

• Published in: European journal of radiology Vol. 192; p. 112380
• Main Authors: Lee, Hyeyun, Park, Gwisang, Park, Chankue, Ok, Jiyun, Lee, Whal, Kim, Sung Mok, Wu, Ming-Ting, Choo, Ki Seok
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.11.2025
• Subjects: Aged; Aged, 80 and over; Algorithms; Aorta - diagnostic imaging; Aorta - pathology; Aortic Aneurysm; Ascending Aorta; Chest Radiograph; Deep Learning; Dilatation, Pathologic - diagnostic imaging; Female; Humans; Male; Middle Aged; Radiographic Image Interpretation, Computer-Assisted - methods; Radiography, Thoracic - methods; Reproducibility of Results; Retrospective Studies; Sensitivity and Specificity; Tomography, X-Ray Computed - methods; CT; AAD; Chest Radiograph; Aortic Aneurysm; AI; ROC AUC; PA CXR; CXR; Deep Learning; Ascending Aorta
• ISSN: 0720-048X; 1872-7727; 1872-7727
• DOI: 10.1016/j.ejrad.2025.112380

•Many AI tools detect ascending aortic dilatation on CT or MR, but few use only a single PA view chest X-ray for detection.•The deep learning algorithm had consistent results (AUC, 0.82-89) across different clinical thresholds and cohorts.•The AI algorithm showed higher sensitivity and lower specificity than human readers, with superior overall performance. This study externally tests the performance of an artificial intelligence algorithm (AI) for diagnosing ascending aortic dilatation (AAD) using PA view chest radiography (PA CXR). Two retrospectively collected cohorts with paired CXR/CT within 30 days (Group 1) and 90 days (Group 2) were gathered as external test sets. The performance of AI (DeepCatch X Aorta v1.1.0) for detecting AAD using PA CXR was analyzed against semi-automatic measurements on CT using fixed thresholds (4.0, 4.5, and 5.0 cm) and subjective assessment (dilatation or no dilatation) of 2 cardiovascular radiologists (>20 years’ experience) and 3 less-experienced (<2 years’ experience) doctors. Additionally, modified thresholds accounting for radiographic magnification were analyzed in the Group 1. Performance metrics (sensitivity, specificity, and AUC) and inter-reader agreement were used. Group 1 included 336 scans (mean age = 64 years) and Group 2 had 190 scans (mean age = 70 years). Dilatation (≥4 cm) was present in 182 patients (Group 1) and 94 patients (Group 2). The algorithm demonstrated robust performance in both Groups for the ≥4 cm threshold, with AUC of 0.89 (95 % CI: 0.85, 0.92) in Group 1 and 0.88 (95 % CI: 0.83, 0.92) in Group 2. All readers’ performances were under the AUC of AI at 4 cm threshold. Inter-reader agreement was fair to moderate (Cohen’s kappa, Group 1, 0.38; Fleiss's kappa, Group 2, 0.37). The AI algorithm detected AAD from a single PA CXR with robust AUC over 0.8 across different clinical thresholds and cohorts, outperforming human readers.