CNSC-26. Quantitative Imaging Analysis for Preoperative Prediction of Cerebellar Mutism Syndrome: A Retrospective Study

OBJECTIVE Cerebellar mutism is a serious complication of posterior fossa surgery. At present, there are few studies on preoperative prediction of cerebellar mutism and a relatively authoritative preoperative evaluation system is lacking. The purpose of this study is to explore the anatomic and pathological mechanism of cerebellar mutism, and establish a pre-operative prediction model of CMS(cerebellar mutism syndrome) based on pre-operative images, which is used to predict the risk probability of CMS after surgery. METHODS The preoperative images and postoperative mutism of 73 patients undergoing posterior fossa surgery in Guangdong Sanjiu Brain Hospital from January 2016 to February 2023 were reviewed. The pre-operative physiological and anatomical risk factors of CMS were determined by quantitative imaging indexes and the pre-operative CMS risk factors were analyzed by Logistic regression. RESULTS A retrospective analysis of pre-operative imaging anatomy of tumor masses revealed that the incidence of tumor invasion to the middle cerebellar peduncle, superior cerebellar peduncle and brainstem in patients with post-operative CMS was significantly higher than in patients without post-operative CMS (p=0.001). The image quantization indexes defined in this study, including coronal index CORd, sagittal index SAGa, axial index OAXa,OAXd, and comprehensive index OAXd*CORd, were statistically significant for the occurrence of CMS after surgery. The results of the binary logistic regression model showed that the accuracy rate of the predictive model in predicting the non-occurrence of CMS after operation was 98.3%, the accuracy rate of predicting the occurrence of CMS after operation was 69.2%, and the overall predictive accuracy rate was 93.2%(68/73). CONCLUSION The invasion of space-occupying objects in the cranial fossa into the middle cerebellar peduncle and superior cerebellar peduncle is a key anatomic risk factor for CMS postoperatively. The quantitative and comprehensive indexes of preoperative images can accurately predict the occurrence of CMS postoperatively, and the prediction model can be used to guide postoperative rehabilitation.