MIT's xvr aligns surgical X-rays with 3D scans in seconds at sub-millimeter precision

MIT News AI · rss · 2026-09-16

MIT CSAIL researchers and collaborators published xvr (X-ray volume registration) in Nature, a technique that rapidly and accurately matches real-time X-rays taken during minimally invasive surgery with preoperative CT/MRI scans, helping clinicians precisely steer catheters and endoscopes.

How it works:

Validation: Tested on the largest real 2D/3D registration dataset spanning five hospitals, dozens of bone and organ systems, and adult plus pediatric patients, xvr outperformed existing AI methods by an order of magnitude while running fast enough for emergency procedures like stroke interventions. The team is now working with surgical robotics companies on navigation tools. Lead author is Vivek Gopalakrishnan; funding came from NIH and other MIT programs.

Original post →

More from Embodied

Embodied channel →