A Framework for multimodal image guidance for robot-assisted partial nephrectomy

Robot-assisted laparoscopic partial nephrectomies (RALPN) are performed to treat patients with locally confined renal carcinoma. However, there are challenges in identifying tumor margins and critical benign structures in this time-critical procedure. The primary objective of this effort is to couple multiple image and data streams together to augment visual information currently provided to surgeons performing RALPN and ultimately ensure complete tumor resection and minimal damage to functional structures. Our framework involves registering high-resolution anatomic images with intra-operative ultrasound (US) and optical-based surface mapping using EM tracking. We have currently implemented methods for surface reconstructions from MR, US, and the stereo-endoscope, US-to- MR registration, and deformation registration using control points. These components have been evaluated through phantom studies to demonstrate protocol feasibility.