Mapping Resection Progress by Tool-Tip Tracking during Brain Tumor Surgery for Real-Time Estimation of Residual Tumor

SIMPLE SUMMARY: Surgical resection continues to be the primary therapeutic strategy in neurosurgical oncology. Computerized cranial neuronavigation based on preoperative imaging can offer precision guidance during early tumor resection but loses validity as the procedure progresses with tissue removal and shifting. Modalities such as intraoperative MRI (iMRI) and intraoperative ultrasound (iUS) can restore image guidance to maximize the extent of resection but present challenges in terms of temporal and spatial resolution, respectively. Our study leverages an untapped data stream from clinical neuronavigation systems to track time-stamped tool-tip positions of surgical instruments. This enables the mapping of resection progress with temporal and spatial accuracy for the real-time estimation of residual tumors. By itself, our technique could serve as an alternative to iMRI for resource-limited regions of the world and as an educational training and evaluation tool. It could also be combined with other intraoperative imaging modalities, such as iUS, to more accurately model and compensate for brain shift. ABSTRACT: Surgical resection continues to be the primary initial therapeutic strategy in the treatment of patients with brain tumors. Computerized cranial neuronavigation based on preoperative imaging offers precision guidance during craniotomy and early tumor resection but progressively loses validity with brain shift. Intraoperative MRI (iMRI) and intraoperative ultrasound (iUS) can update the imaging used for guidance and navigation but are limited in terms of temporal and spatial resolution, respectively. We present a system that uses time-stamped tool-tip positions of surgical instruments to generate a map of resection progress with high spatial and temporal accuracy. We evaluate this system and present results from 80 cranial tumor resections. Regions of the preoperative tumor segmentation that are covered by the resection map (True Positive Tracking) and regions of the preoperative tumor segmentation not covered by the resection map (True Negative Tracking) are determined for each case. We compare True Negative Tracking, which estimates the residual tumor, with the actual residual tumor identified using iMRI. We discuss factors that can cause False Positive Tracking and False Negative Tracking, which underestimate and overestimate the residual tumor, respectively. Our method provides good estimates of the residual tumor when there is minimal brain shift, and line-of-sight is maintained. When these conditions are not met, surgeons report that it is still useful for identifying regions of potential residual.