Key result
VB-Net deep learning model detects aortic dissection with ~0.99 AUC while markedly slashing reconstruction time.
Why the study?
Aortic dissection is a life-threatening cardiovascular emergency requiring rapid and accurate diagnosis, motivating the clinical validation of a VB-Net-based deep learning model for automated segmentation, detection, and Stanford classification on computed tomography angiography.
Does a VB-Net-based deep learning model accurately detect and classify aortic dissection on aortic CTA images and reduce processing time compared to expert radiologist adjudication and manual reconstruction?
Observational (n=378)
Double-blind
Yes
Does a VB-Net-based deep learning model accurately detect and classify aortic dissection on aortic CTA images and reduce processing time compared to expert radiologist adjudication and manual reconstruction?
This multicenter validation demonstrates that a VB-Net deep learning model can accurately detect and classify aortic dissection on CTA with near-perfect agreement to expert radiologists while reducing 3D reconstruction time from over 10 minutes to 33 seconds.
Answers are grounded in this paper and cite their sources.
Supports automated CTA detection in retrospective data; leaves open prospective validation before clinical adoption.
Background Aortic dissection (AD) is a life-threatening cardiovascular emergency requiring rapid and accurate diagnosis. This study aimed to clinically validate a VB-Net-based deep learning model for automated segmentation, detection and Stanford classification of AD on computed tomography angiography (CTA). Methods This multicenter retrospective study enrolled 378 patients who underwent aortic CTA between July 2022 and September 2023 across three independent institutions. Golden standard was established by two senior radiologists, with discrepancies adjudicated by a third expert radiologist (≥ 25 years of experience). Model performance for AD diagnosis and Stanford classification was evaluated using area under the curve (AUC), sensitivity, specificity, and Cohen’s Kappa statistics. Results The inter-reader agreement between the two senior radiologists was substantial for both AD presence and Stanford classification (kappa: 0.96 and 0.99, respectively). The model demonstrated excellent diagnostic agreement across centers. For AD detection, Cohen’s kappa values were 0.99, 0.98, and 0.96, with AUCs of 1.00 (95% CI: 0.98–1.00), 1.00 (0.96–1.00), and 0.99 (0.95–1.00) at Centers 1–3, respectively. Sensitivity and specificity ranged from 89.31% to 95.63% and 88.46% to 93.56%, with all 95% CI lower bounds exceeding 80%. For Stanford classification, kappa values were 0.93–0.96, with sensitivity and specificity consistently above 82%. Notably, AI-driven reconstruction reduced processing time from 10 min 18 s (manual) to 33.27 s per case (P < 0.001). Conclusion The model enables precise and rapid AD detection and Stanford classification using aortic CTA images, with high consistency and efficiency across multiple centers, demonstrating strong clinical applicability.
No takes yet. Share an insight, caveat, or question.
Song et al. (2026) conducted an observational in Aortic dissection (n=378). VB-Net-based deep learning model for automated aortic dissection diagnosis and Stanford classification vs. Gold standard consensus of senior radiologists and manual 3D reconstruction was evaluated on Diagnostic accuracy for aortic dissection detection (AUC, sensitivity, specificity, Cohen's kappa) and reconstruction processing time (AUC 0.99-1.00; Cohen's kappa 0.96-0.99; Sensitivity 97.96%-100.00%; Specificity 96.61%-100.00%, 95% CI 0.95-1.00, p=<0.001). The VB-Net deep learning model achieved AUCs of 0.99-1.00 for aortic dissection detection on CTA and reduced reconstruction time from 618 s to 33.27 s (P<0.001).
Synapse has enriched 10 related papers. Consider them for comparative context: