AI Insight
Researchers fine-tuned Meta's SAM2 segmentation model to identify coronary arteries in X-ray fluoroscopy images, addressing challenges like low contrast and imaging noise. The fine-tuned model achieved a Dice score of 0.767 on the ARCADE validation dataset (200 images), compared to 0.033 for the zero-shot baseline, demonstrating substantial improvement. Testing on 10 fluoroscopic video studies showed the model could track vessels accurately and distinguish them from other structures like ribs and stents in 9 of 10 cases.
Why it matters
Accurate automated segmentation of coronary arteries in real-time fluoroscopy could assist cardiologists during interventional procedures, potentially improving procedural efficiency and reducing interpretation errors. The released code and model checkpoint enable other researchers and clinicians to implement and build upon this work.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
SAM2 (Meta, 2024) provides a strong starting point for segmentation, but given the unique challenges in medical imaging (noise from patient movement, the projection-based nature of X-ray fluoroscopy, and low contrast between vessels and background), direct application is difficult. We fine-tune MedSAM2 on annotated coronary angiograms and apply it to video data for point-of-care use. On the ARCADE validation set (200 images), the fine-tuned model achieves Dice 0.767 compared to 0.033 zero-shot. On 10 fluoroscopic video studies from CoronaryDominance, it tracks vessels coherently and avoids falsely segmenting ribs, stents, and bypass grafts in 9 of 10 studies. Code is available at https://github.com/elakiyasivakumar/SAM2-Coronary-Angiography-VA and the fine-tuned checkpoint at https://huggingface.co/Elakiya17/CA-SAM2.
Source: Fine-Tuning SAM2 for Coronary Artery Segmentation in X-Ray Fluoroscopy