MASAMUNE Ken
Department Graduate School of Medical Science, Graduate School of Medical Science Position Professor |
|
Language | Japanese |
Title | A study on Surgical Process and Procedure(or technique)Identification Method Using
Pre‒ and Intra‒operative Information in Awake Surgery of Glioma |
Conference | The 28th Annual Congress of Japan Society of Computer Aided Surgery |
Conference Type | Nationwide Conferences |
Presentation Type | Speech |
Lecture Type | General |
Date | 2019/11/24 |
Society abstract | 日本コンピュータ外科学会誌 21(4),339-340 2019 |
Summary | Abstract:During surgery for glioma, surgeon will resect a glioma based on experience and technique, by considering
tumor position, brain function, and surrounding structure. The experience and technique of surgeon is implicit knowledge; the visualization of this knowledge is important for improvement and educational support to increase the quality of medical care. Therefore, we aim to visualize the knowledge by the identification and analysis of surgical processes, based on previous research1). As a result, Visualization of implicit knowledge requires refinement of the model and visualization of the procedure. In this study, we propose of surgical processes model and identification method process using machine learning in awake surgery. First, we subdivided the process that includes the surgeonʼs procedure into 13 processes as the fourth layer in the previous 3 layered HHMM model. Seconds, we automatically extract the features using machine learning, image and signal processing. Finally, surgical processes are identified using the new features. The method has been evaluated on the past data. Accuracy of the identified processes was 63.2%. Feature extraction accuracy greatly affects identification accuracy. Therefore, more accurate extraction is necessary in the future. Our future work is development of surgical process analysis system to visualize of implicit surgical knowledge. Key words:Surgical process, Machine Learning, Modeling |