See, Plan, Cut: MPC-Based Autonomous Volumetric Robotic Laser Surgery with OCT Guidance
Ravi Prakash, Vincent Wang, Arpit Mishra, Devi Yuliarti, Pei Zhong, Ryan McNabb, Patrick Codd, Leila Bridgeman
AI summary
Problem
Current robotic laser surgery systems lack integrated volumetric sensing, precise imaging-to-laser calibration, and constraint-aware planning, forcing reliance on open-loop or simulation-only methods that fail in complex 3D surgical environments.
Approach
The authors developed RATS, an OCT-guided robotic platform that integrates high-resolution imaging with a fiber-coupled laser, using a super-Gaussian laser-tissue interaction model and a sampling-based model predictive control framework to generate closed-loop, constraint-aware resection trajectories.
Key results
- Achieved 0.161 mm OCT-to-laser calibration accuracy
- Super-Gaussian LTI model outperforms Gaussian baselines with 0.231 mm RMSE
- MPC planner reduces resection error to 0.842 mm and improves IoU by 64.8% over feedforward execution
- First demonstration of closed-loop autonomous volumetric laser resection with subsurface structure preservation
Why it matters
Enables precise, obstacle-aware soft tissue removal with direct clinical potential for neurosurgery and complex tumor resections.
Abstract
Robotic laser systems enable sub-millimeter, non- contact tissue resection, yet existing platforms lack volumet- ric planning and intraoperative feedback. We present RATS (Robot-Assisted Tissue Surgery), an intelligent optical co- herence tomography (OCT)-guided robotic platform for au- tonomous volumetric soft tissue resection. RATS integrates macro-scale RGB-D imaging, micro-scale OCT, and a fiber- coupled surgical laser, calibrated through a novel multistage alignment pipeline that achieves OCT-to-laser calibration ac- curacy of 0.161 ± 0.031 mm. A super-Gaussian laser–tissue interaction (LTI) model characterizes ablation morphology with an average RMSE of 0.231 ± 0.121 mm, outperform- ing Gaussian baselines. A sampling-based model predictive control (MPC) framework operates directly on OCT voxel data to generate closed-loop, constraint-aware resection trajec- tories, achieving 0.842 mm RMSE (root-mean-square error) and improving intersection-over-union agreement by 64.8% compared to feedforward execution. RATS also detects and preserves subsurface structures, demonstrating the first closed- loop autonomous volumetric robotic laser resection with OCT guidance. To our knowledge, this is the first demonstration of closed-loop autonomous volumetric robotic laser resection with OCT guidance, enabling precise, obstacle-aware tissue removal with potential in neurosurgery.