Back

Fine-Tuning SAM2 for Coronary Artery Segmentation in X-Ray Fluoroscopy

Sivakumar, E.

2026-06-19 radiology and imaging
10.64898/2026.06.16.26355803 medRxiv
Show abstract

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.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

1
Medical Image Analysis
○ 35 papers in training set
Top 0.1%
9.4%
Similar papers in this journal
2
Journal of Medical Imaging
● 11 papers in training set
Top 0.1%
9.4%
Similar papers in this journal
3
Nature Communications
○ 5641 papers in training set
Top 24%
6.6%
Similar papers in this journal
4
Nature Machine Intelligence
○ 70 papers in training set
Top 0.4%
6.1%
Similar papers in this journal
5
Scientific Reports
○ 3612 papers in training set
Top 13%
6.1%
Similar papers in this journal
6
Medical Physics
○ 14 papers in training set
Top 0.1%
5.3%
7
IEEE Access
● 35 papers in training set
Top 0.3%
3.9%
8
PLOS ONE
● 5266 papers in training set
Top 35%
3.3%
50% of probability mass above
9
IEEE Transactions on Medical Imaging
● 21 papers in training set
Top 0.2%
3.1%
10
Frontiers in Computational Neuroscience
○ 60 papers in training set
Top 0.4%
3.1%
11
PLOS Computational Biology
● 1863 papers in training set
Top 12%
2.4%
12
Journal of the American Medical Informatics Association
◐ 71 papers in training set
Top 1%
2.3%
13
GigaScience
◐ 212 papers in training set
Top 2%
2.3%
14
European Heart Journal - Digital Health
◐ 18 papers in training set
Top 0.4%
2.1%
15
npj Digital Medicine
○ 118 papers in training set
Top 2%
1.7%
16
Imaging Neuroscience
● 282 papers in training set
Top 3%
1.7%
17
Patterns
○ 78 papers in training set
Top 1%
1.6%
18
PLOS Digital Health
● 106 papers in training set
Top 3%
1.4%
19
NeuroImage
○ 903 papers in training set
Top 5%
1.4%
20
Computers in Biology and Medicine
○ 128 papers in training set
Top 3%
1.4%
21
Nature Methods
○ 385 papers in training set
Top 5%
1.3%
22
Frontiers in Neuroinformatics
○ 41 papers in training set
Top 0.4%
1.3%
23
IEEE/ACM Transactions on Computational Biology and Bioinformatics
● 38 papers in training set
Top 0.7%
1.3%
24
Expert Systems with Applications
○ 11 papers in training set
Top 0.2%
1.3%
25
European Radiology
○ 15 papers in training set
Top 0.4%
1.3%
26
Brain Informatics
○ 10 papers in training set
Top 0.2%
1.1%
27
Nature
○ 645 papers in training set
Top 9%
1.0%
28
BMC Bioinformatics
○ 457 papers in training set
Top 5%
0.9%
29
Frontiers in Plant Science
○ 256 papers in training set
Top 4%
0.8%
30
BMC Medical Informatics and Decision Making
○ 43 papers in training set
Top 2%
0.8%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.