P3‐347: INDIVIDUAL EVALUATION SYSTEM FOR WHITE MATTER HYPERINTENSITY RECOGNITION USING DEEP CONVOLUTIONAL NEURAL NETWORK
Jin San Lee, Kyung Mi Lee, Eui Jong Kim, Hak Young Rhee, Key-Chung Park, Jang-Hoon Oh, Jae‐Ho Lee, Hyun Sub Lee, Hyug-Gi Kim · Alzheimer s & Dementia · 2018
White matter hyperintensities (WMH) is one of the main imaging feature to evaluate cerebral small vessel disease (cSVD). The objective of this study to investigate the feasibility of WMH recognition using deep convolutional neural networks (CNN). Furthermore, individual evaluation system was proposed to classify WHM groups. A total of 500 elderly healthy subjects were participated after informed consent. All subjects were scanned on the 3T MR with 2D-axial FLAIR sequences. The WMH score was labeled four groups- none, mild, moderate and severe. Input dataset was preprocessed for data augmentation (Readbrain Co. Ltd). We implemented on the data to differentiate among WHM groups with the AlexNet CNN model that is a powerful deep learning architecture for imaging classification. The results of classification were quantitatively assessed by accuracy and evaluated using the activated feature maps in each layer. Two radiologists were reviewed all classified images to validate labeling. WMH groups were classified from result the review of two neuroradiologists (kappa= 0.8). The accuracy of the set of training data with deep CNN model was 96.06 % and the result of classification with the set of test data showed that the accuracy was 91.67% for four labeled groups. Learned WMH classifier by deep learning recognized high intensity as feature of WMH. These activated features of the WMH were visually confirmed to be closely correlated with the features evaluated FLAIR MR image reading by radiologists. The deep CNN based on personalized WMH evaluation system proved to be more effective to evaluate the degree of WMH process. Therefore, the deep learning based on WMH evaluation method can be used as an adjunct to diagnosis cSVD.