10237- MNG-8 AUTOMATED VOLUMETRY OF MENINGIOMAS IN CONTRAST-ENHANCED T1-WEIGHTED MRI USING DEEP LEARNING

Abstract BACKGROUND With the widespread use of MRI equipment and brain docks, there are more opportunities to perform follow-up observation for meningiomas. In recent years, attempts have been made to predict prognosis by using tumor volume as an imaging biomarker. On the other hand, manual lesion extraction by a reader with anatomical knowledge is indispensable for volume measurement, which requires a certain amount of labor and time. We developed a simple and highly reproducible application for automated volume measurement of meningiomas using artificial intelligence and report its accuracy for use as an imaging biomarker of tumor volume in real clinical practice. METHODS 178 patients with meningiomas who underwent MRI at Osaka University Hospital were included in the study. Contrast-enhanced T1-weighted images were used for manual lesion extraction by two neurosurgeons. These annotated image data were randomly divided into training (90%) and testing (10%). The deeplabV3 model was used to train the artificial intelligence, and leave-one-out cross-validation was used to train and validate the automated model. The test data were used for comparison, and the agreement rate (Dice index) on the images of the automatically extracted regions and the error rate of the calculated tumor volume (mean square error rate: RMSPE) were used as indices of measurement performance. Results: The agreement rate between the images of the automatically extracted area and the test data was 92.3± 5.1% on average. The error rate for volume was 9.5±1.2%. Conclusion: This application using artificial intelligence was considered to have a certain validity in terms of accuracy of automatic lesion extraction.