A new preprint led by Elakiya Sivakumar at Columbia University reports a targeted fix for detecting fine arteries on moving X-rays during coronary angiograms. By fine-tuning MedSAM2 for coronary angiograms, the system improved from roughly 3 out of 100 to roughly 77 out of 100 on a vessel-overlap score in 200 validation images, then tracked arteries consistently in most test videos. The ARCADE challenge was created for this bottleneck, as manually selecting coronary segments and stenotic lesions is costly and time-consuming. The new work starts with Meta’s Segment Anything Model 2, using MedSAM2, a version adapted for medical images and video built by Jun Ma, Bo Wang, and collaborators. Instead of asking the model to solve a new hospital task cold, Sivakumar trained it on coronary angiograms. The model was trained and tested with the ARCADE coronary angiography data; the public Phase 1 set includes 1,200 vessel-tree images, split into 1,000 training images and 200 validation images, with annotations. The GitHub record says the team used a single click near the center of the target vessel as a prompt, and training expanded the examples through rotations and flips.