Real-time, high-precision detection of aortic dissection key components for emergency triage using an enhanced deep learning model
Related Experiment Video
Updated: May 12, 2026
An Approach to Point-Of-Care Ultrasound Evaluation of the Abdominal Aorta
Published on: September 8, 2023
Houde Wu1, Fei Chen2,3,4,5, Ruiqing Xia6
1School of Medical Technology, Tianjin Medical University, Tianjin, China.
Insights
A new deep learning model, AD-YOLO11, rapidly and accurately identifies key components of Type B aortic dissection (TBAD) from CTA scans. This tool aids in emergency triage by providing real-time analysis for critical cardiovascular conditions.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Cardiovascular Surgery
Background:
- Aortic dissection (AD) is a critical cardiovascular emergency requiring precise imaging analysis.
- Accurate identification of true lumen (TL), false lumen (FL), and false lumen thrombus (FLT) in Type B AD (TBAD) is vital for patient management.
- Existing deep learning methods struggle with real-time detection and FLT identification in TBAD.
Purpose of the Study:
- To develop a real-time, high-precision deep learning framework for simultaneous detection of TL, FL, and FLT in TBAD.
- To enhance emergency triage through rapid and accurate cardiovascular imaging analysis.
Main Methods:
- An enhanced YOLOv11 model, AD-YOLO11, was developed.
- Key innovations include Recursive Information Distillation Network (RIDNet) for noise suppression, Triplet Attention Mechanism for feature enhancement, and MPDIoU loss for bounding box regression.
- The model was trained and validated on extensive datasets of CT angiography (CTA) slices from TBAD patients.
Main Results:
- AD-YOLO11 demonstrated high precision (0.991) and recall (0.936) on internal validation.
- The model achieved excellent mean Average Precision (mAP@0.5:0.95 of 0.883) and maintained strong generalizability on external validation.
- Achieved millisecond-level inference speeds on GPU (3.18 ms/slice) and clinical feasibility on CPU (53.15 ms/slice).
Conclusions:
- AD-YOLO11 provides high-accuracy, millisecond-level detection of critical TBAD components from CTA.
- Its efficient performance on both GPU and CPU positions it as a valuable tool for emergency triage and resource-limited settings.
- This deep learning framework complements traditional 3D segmentation methods for faster aortic dissection assessment.
Background:
Aortic dissection (AD) is a life-threatening cardiovascular emergency. For Type B AD (TBAD), rapid and accurate identification of the true lumen (TL), false lumen (FL), and false lumen thrombus (FLT) from CTA is critical for risk stratification and treatment planning. However, existing deep learning solutions often lack real-time capability and fail to address the detection of FLT.
Purpose:
To develop a real-time, high-precision deep learning framework for the simultaneous detection of all three key AD components to support emergency triage.
Methods:
We propose AD-YOLO11, an enhanced YOLOv11 model integrating three key innovations: (1) a Recursive Information Distillation Network (RIDNet) for CTA noise suppression, (2) a Triplet Attention Mechanism for spatial and channel feature enhancement, and (3) the MPDIoU loss function for optimized bounding box regression. The model was trained and internally validated on a dataset of 25 176 slices from 106 TBAD patients and externally validated on 18 238 slices from 71 independent patients.
Results:
On internal validation, AD-YOLO11 achieved a precision of 0.991 ± 0.004, recall of 0.936 ± 0.006, mAP@0.5 of 0.951 ± 0.007, and mAP@0.5:0.95 of 0.883 ± 0.008. It maintained high performance on the external test set, demonstrating strong generalizability. The inference speed was 3.18 ± 0.23 ms per slice on GPU, and it remained clinically feasible on CPU (53.15 ± 2.76 ms per slice).
Conclusions:
AD-YOLO11 achieves millisecond-level, high-accuracy detection of all three critical Type B aortic dissection components from CTA images. Its efficient inference on both GPU and CPU makes it a promising frontline tool for rapid triage in emergency and resource-limited settings, effectively complementing time-consuming 3D segmentation for aortic dissection assessment.
Related Concept Videos
01:22Aortic Regurgitation II: Clinical Features and Diagnostic Tests
01:21Aneurysm II: Clinical Manifestations and Diagnostic Studies