New criteria for histopathological classification of testis based on Johnsen score for male infertility using automated deep learning software

Background: Johnsen scores are commonly used for scoring testicular biopsy, and are needed for testicular sperm extraction (TESE). However, pathologists need much experience to make histopathological evaluations of testes. Therefore, we considered that a tool for determining Johnsen scores automatically using AI could be used in place of traditional Johnsen scoring to support pathologists’ evaluations. Methods: 275 patients with obstructive or non-obstructive azoospermia who underwent testicular sperm extraction (TESE) were included in this study. We obtained testicular tissues for the 275 patients and were able to make 264 haematoxylin and eosin (H&E)-stained glass microscope slides. Slides were examined under an optical microscope (X400). In addition, we cut out of parts of the histopathology images of (5·0 X 5·0 cm) for expansion of Johnsen’s characteristic areas with seminiferous tubules. We defined four labels: Johnsen score 1-3, 4-5, 6-7, and 8-10 to distinguish Johnsen scores in clinical practice. All images were uploaded to the Google Cloud AutoML vision platform. We obtained a dataset of 7155 images at magnification X400 and a dataset of 9822 expansion images for the 5·0 X 5·0 cm cutouts. Performance of the algorithm was assessed by average precision (positive predictive value), precision, and recall. Findings: For the X400 magnification image dataset, the average precision (positive predictive value) of the algorithm was 82·6%, precision was 80·31%, and recall was 60·96%. For the expansion image dataset (5·0 X 5·0 cm), the average precision of the algorithm was 99·5%, precision was 96·29%, and recall was 96·23%. Interpretation: This is the first report of an AI-based algorithm for predicting Johnsen scores. It was created through an automated deep learning approach needing no coding experience. Funding: Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science Declaration of Interests: All authors have no competing interests to declare. Ethics Approval Statement: The study was approved by the Research Ethics Committee of the Omori Hospital, School of Medicine, Toho University (approval No. M20103).