Robot-assisted intraoperative ultrasound for spine image-guidance

The National Institute of Arthritis and Musculoskeletal and Skin Diseases awarded Worcester Polytechnic Institute $368,897 on July 15, 2026, under the Arthritis, Musculoskeletal and Skin Diseases Research program (CFDA 93.846) to develop robot-assisted intraoperative ultrasound registration techniques for spinal image-guidance in surgical procedures.

WPI will develop an automatic, level-wise ultrasound-CT registration pipeline to enable ultrasound-based image-guidance as an alternative to intraoperative CT in both minimally-invasive and open spinal procedures. The team will evaluate registration effectiveness using explanted porcine spines and live pigs, establishing ground-truth registration through intraoperative CT and mini-screw implants in exposed vertebrae, then comparing relative performance across accuracy, efficiency, and robustness metrics to assess clinical feasibility. WPI will leverage newly established capabilities at its institution and the UMass Chan Medical School for intraoperative navigation, imaging, and intraoperative CT access to deliver a functioning image-guidance system for spinal procedures by project completion.

Performance occurs in Worcester, Massachusetts. The award runs through June 30, 2028.

Abstract

Image-guidance has not been widely adopted in open spinal procedures, despite the fact that IT improves surgical accuracy and patient outcomes. This is largely because of limitations in patient registration and poor reputation as a direct consequence rather than procedure-specific translation of technology from open cranial applications. We propose to develop patient registration techniques using robot-assisted intraoperative ultrasound to establish accurate, reliable and robust patient registration with respect to preoperative CT images of the spine. There are two specific aims. First, we will develop an automatic and level-wise US-CT registration pipeline so that to enable the use of ultrasound for image-guidance for both minimally-invasive and open spinal procedures as an alternative to intraoperative CT. Second, we will evaluate the effectiveness of these registration schemes using explanted porcine spines and live pigs. The "ground-truth" registration will be established using intraoperative CT and mini screws implanted to exposed vertebrae. Through these efforts, we will compare the relative performances (accuracy, efficiency and robustness) of different patient registration techniques to assess their relative clinical feasibility for practical applications. The proposed research will be facilitated by the newly established practicepoint at our institution as well as those by the umass chan medical school for intraoperative navigation and imaging capabilities including intraoperative CT (ICT). We anticipate that by the end of the proposed program, we will create a functioning image-guidance system for spinal procedures to gain an initial understanding of its effectiveness in actual clinical applications.