Medical Robotics (1)

2020 Biomedical Engineering  
Introduction Surgical microscopes in neurosurgery are used for magnified visualization of the situs during e.g. brain tumor removal or aneurysm surgery. Due to dynamic surgical events, such as tissue removal, surgeons frequently move the microscope manually to maintain a sufficient view throughout surgery. Development of user-friendly surgical microscopes requires an in-depth understanding of the surgeon's interaction with the microscope. Previous work focused on the question "when" surgeons
more » ... eracted with the microscope, by quantifying interaction frequencies based on video analysis. Our work aims at extending prior approaches to "how" surgeons interact with the surgical microscope by identifying motion behaviour patterns. Methods Our hypothesis is that different motion type behaviours, such as fine or coarse motions, can be distinguished from surgical data. In a first step, we conducted a qualitative analysis by observing various neurosurgical surgeries in a large neurosurgical clinic. For a quantitative analysis we use motion data from each of ten tumor and vascular cases visualized by a ZEISS KINEVO 900. We are currently evaluating different spatio-temporal data mining methods employing trajectory descriptors, such as Frenet-Serret invariants. Based on different motion types identified in the data, we are comparing characteristic motion patterns between tumor and vascular cases. Results First qualitative observations indicate that motions of surgical microscopes in neurosurgery can be divided in fine translational and pivoting motions, as well as coarse motions. In the tumor cases, qualitatively, small motions are more frequent, whereas in the vascular cases motions are larger in magnitude. These observations serve as a baseline for later quantitative comparison. Conclusion Our work supports the development of user-friendly surgical microscopes by analysing their intraoperative motion patterns during neurosurgical interventions. Future work includes validation of the quantified motion patterns derived from recorded data and quantitative comparison of motion types in tumor and vascular cases.
doi:10.1515/bmt-2020-6013 pmid:33578476 fatcat:yulkzep5mbhjzkvst7gidoudxa