Deep Learning–Based Automated Imaging Classification of ADPKD

Abstract

IntroductionThe Mayo imaging classification model (MICM) requires a pre-step qualitative assessment to determine whether a patient is in class 1 (typical) or class 2 (atypical), where patients assigned to class 2 are excluded from the MICM application. MethodsWe developed a deep learning-based method to automatically classify class 1 and 2 from MR images and provide classification confidence utilizing abdominal T2-weighted magnetic resonance images from 486 subjects, where the transfer learning was applied. In addition, the explainable artificial intelligence (XAI) method is illustrated to enhance the explainability of the automated classification results. For performance evaluations, the confusion matrices were generated and receiver operating characteristic curves were drawn to measure the area under the curve. ResultsThe proposed method showed excellent performance for the classification of class 1 (97.7%) and 2 (100%), where the combined test accuracy of 98.01%. The precision and recall for predicting the class 1 were 1.00 and 0.98, respectively, with the F1-score of 0.99, while those for predicting the class 2 were 0.87 and 1.00, respectively, with the F1-score of 0.93. The weighted averages of precision and recall were 0.98 and 0.98, respectively, showing the classification confidence scores while the XAI method well highlighted contributing regions for the classification. ConclusionsThe proposed automated method can classify the class 1 and 2 cases as accurately as the level of a human expert, which may be a useful tool to facilitate clinical trials investigating different types of kidney morphology and clinical management of patients with ADPKD.