297 Impact of Navigated Intraoperative Ultrasound on Brain Shift and Gross Tumor Resection

Abstract

INTRODUCTION: Intraoperative imaging and neuronavigation aid in achieving maximum tumor resection. However, tissue removal during surgery can cause brain shift, which alters the relative anatomy and “de-calibrates” the standard approaches. While intraoperative MRI (iMRI) is the gold-standard to correct for brain shift, it is expensive, time-consuming, and logistically difficult. As an alternative, we investigated an approach which combines intraoperative ultrasound (ioUS) technology with preoperative imaging to allow for real-time compensation of brain shift and assessment of extent of resection (EOR). METHODS: We prospectively enrolled patients undergoing neurosurgical tumor resection at Massachusetts General Hospital beginning April 2023. Patients were registered with Brainlab Ultrasound Navigation prior to surgery. ioUS sweeps were acquired before cortex incision, mid-resection, and post-resection with bkActiv and fused to preoperative MRIs. Four points around the tumor were longitudinally tracked to assess brain shift. Data included sweep duration, image processing speed, extent of brain shift as determined by intraoperative markers, and EOR assessed from the ioUS compared with postoperative MRIs. RESULTS: Seven patients were enrolled (average age 62 ± 15, 75% male) with frontal (n = 5), temporal (n = 1), and occipital (n = 1) gliomas (n = 4) and metastases (n = 3). All patients exhibiting total resection as determined by ioUS also demonstrated total resection on post-operative MRI, evaluated independently by a neuroradiologist. The average time of scan and image processing (n = 19) was 5.7 ± 1.9 and 27.8 ± 6.5 seconds respectively. CONCLUSIONS: Navigated ioUS offers accurate and efficient assessment of EOR during surgery for tumor resections. Furthermore, when brain shift is detected, navigated ioUS can allow for compensation and continued utilization of navigation via the acquisition of a brand new, navigable, intraoperatively acquired dataset.