Logo

American Heart Association

  80
  0


Final ID: Th0077

Vessel-Specific CT Calcium Scoring in Peripheral Artery Disease Using Deep Learning

Abstract Body: Objective: CT calcium scoring is a tool for assessing disease severity and risk for adverse events in coronary artery disease; however, quantification of vessel-specific calcium burden from CT images in peripheral artery disease (PAD) has remained relatively understudied due to the time-consuming nature of segmenting the arterial network. Therefore, we sought to test the performance of a semi-automated deep learning approach to segment and quantify vessel-specific calcium burden from CT images in PAD patients to streamline eventual clinical implementation of calcium scoring in PAD.
Methods: Patients with PAD (N=80) were prospectively enrolled for non-contrast CT imaging. Images were manually segmented to quantify calcium mass for the femoral-popliteal, peroneal, anterior tibial, and posterior tibial arteries. Manually processed images were used as input data to train an nnU-Net deep learning model. Data augmentation techniques were applied to increase the dataset to 157 images (80 patients=157 legs) to achieve better generalization of results. The dataset was randomly split using an 90/10 ratio for model training and testing. Dice coefficient was calculated to assess the agreement between manual and deep learning image analysis results.
Results: Deep learning-guided image segmentation results qualitatively agreed with manual image analysis (Fig. 1A). Quantitatively, deep learning produced dice coefficients of 0.82 ± 0.05 for femoral-popliteal, 0.71 ± 0.12 for anterior tibial, 0.74 ± 0.12 for posterior tibial, and 0.73 ± 0.12 for the peroneal artery, thus representing good performance for multi-vessel segmentation (Fig. 1B). Calcium mass values derived from both manual and deep learning image analysis demonstrated excellent agreement, with an intraclass correlation coefficient of 0.98 (Fig. 1C).
Conclusions: Deep learning allows for accurate quantification of vessel-specific CT calcium values for the lower extremities of PAD patients, which is a challenging task due to arteries comprising a small percentage (i.e., 0.07%) of the overall CT image. This AI-based approach significantly reduces CT image analysis time from hours to seconds and represents a promising approach for future risk stratification in PAD.
  • Musini, Kumudha  ( The Ohio State University , Columbus , Ohio , United States )
  • Rimmerman, Eleanor  ( The Ohio State University , Columbus , Ohio , United States )
  • Chou, Ting-heng  ( NTUNHS , Taipei , Taiwan )
  • Shin, Kyle  ( The Ohio State University , Columbus , Ohio , United States )
  • Bobbey, Adam  ( Nationwide Children's Hospital , COLUMBUS , Ohio , United States )
  • Atway, Said  ( Ohio State College of Medicine , Columbus , Ohio , United States )
  • Corriere, Matthew  ( Ohio State College of Medicine , Columbus , Ohio , United States )
  • Go, Michael  ( Ohio State College of Medicine , Columbus , Ohio , United States )
  • Stacy, Mitchel  ( Ohio State College of Medicine , Columbus , Ohio , United States )
  • Author Disclosures:
    Kumudha Musini: DO NOT have relevant financial relationships | Eleanor Rimmerman: DO NOT have relevant financial relationships | Ting-Heng Chou: No Answer | Kyle Shin: DO NOT have relevant financial relationships | Adam Bobbey: No Answer | Said Atway: No Answer | Matthew Corriere: No Answer | Michael Go: No Answer | Mitchel Stacy: DO NOT have relevant financial relationships
Meeting Info:
Session Info:

15. Poster Session 3 & Reception

Thursday, 04/24/2025 , 05:00PM - 07:00PM

Poster

More abstracts on this topic:
3D Statistical Shape Analysis Predicts Type A Aortic Dissection Better Than Aortic Diameters

Marway Prabhvir, Campello Jorge Carlos Alberto, Wagner Catherine, Baker Timothy, Burris Nicholas

A KLF2-BMPER-Smad1/5 checkpoint regulates high fluid shear stress-mediated artery remodeling

Deng Hanqiang, Zhang Jiasheng, Schwartz Martin

More abstracts from these authors:
Vessel-by-Vessel CT Calcium Scoring in Peripheral Artery Disease: Association with Patient-Level Factors

Rimmerman Eleanor, Musini Kumudha, Chou Ting-heng, Trainer Jenna, Atway Said, Go Michael, Stacy Mitchel

Deep Learning-Guided CT Image Analysis Quantifies 18-Month Changes in Regional Muscle Atrophy in Patients with Peripheral Artery Disease

Musini Kumudha, Rimmerman Eleanor, Chou Ting-heng, Shin Kyle, Go Michael, Stacy Mitchel

You have to be authorized to contact abstract author. Please, Login
Not Available