Automatic Surgical Workflow Estimation Method for Brain Tumor Resection Using Surgical Navigation Information

It has been acknowledged as a problem in recent years that surgery has become complex due to medical system updating. To respond to the increasing demand for making surgery more optimal and efficient, studies on surgical process analysis have attracted attention. Automatic estimation technology is necessary for accurate and efficient process analysis. With a focus on this problem, we have studied technologies on the automatic estimation of surgical processes. In this study, we develop an automatic estimationmethod for a chosen surgical process on the basis of information obtained from a surgical navigation system, taking as an example image-guided brain tumor surgery. We found a significant correlation among five parameters – progress in enucleation, depth of surgical tool tip, displacement of surgical tool, volume of surgical tool position log data, and number of events detected during surgery – that are defined according to the anatomical information on patients and surgical procedure information on surgeons stored in the navigation system, and three stages in the brain tumor removal process: (1) incision of the surface cortex, (2) testing and blood vessel resection, (3) resection and removal of tumors. By using automatic Bayesian estimation of tumor removal processes in eight case examples using the five parameters, we estimated 73% of all processes correctly. This result indicates that surgical processes are automatically estimated with information in the surgical navigation system alone, which thus contributes to the accurate and efficient surgery analysis.