Kawai, Honami

ABSTRACT Aims Muscle morphology provides important information in differentiating disease aetiology, but its measurement remains challenging because of the lack of an efficient and objective method. This study aimed to quantitatively refine the morphological features of muscle fibres in neuromuscular diseases using machine learning. Methods In this retrospective study, we analysed muscle biopsy specimens on haematoxylin and eosin‐staining. Machine learning–based software was developed to segment muscle fibre contours and perform automated muscle morphometry and subsequent graph theory–based spatial analysis of atrophied fibre grouping. A decision tree–based framework, LightGBM, was trained to predict underlying aetiologies based on morphometric and spatial variables. Results The study included 100 muscle samples, including 20 normal muscles, 49 myopathies and 19 neuropathies. The fine‐tuned segmentation model, YOLOv8, achieved a mask average precision of 0.819. The muscle morphometry revealed the significance of fibre circularity. The mean circularity was higher in the myopathy group, and the SD of circularity was elevated in the neuropathy group. Although most cases were consistent with textbook findings, atypical presentations, such as dermatomyositis with angular atrophy and amyotrophic lateral sclerosis with round atrophy, were objectively documented. Spatial analysis quantified grouped atrophy, showing the potential to feature specific atrophy patterns. The LightGBM model successfully predicted the final clinical diagnosis of the myopathies and neuropathies with an accuracy of 0.852, which exceeded that of 0.808 by human annotation. Conclusion Automated muscle morphometry and spatial analysis provide quantification of muscle morphology and patterns of atrophy, which will facilitate objective and efficient investigation of neuromuscular diseases.

Abstract Aim The morphometry of sural nerve biopsies, such as fibre diameter and myelin thickness, helps us understand the underlying mechanism of peripheral neuropathies. However, in current clinical practice, only a portion of the specimen is measured manually because of its labour‐intensive nature. In this study, we aimed to develop a machine learning‐based application that inputs a whole slide image (WSI) of the biopsied sural nerve and automatically performs morphometric analyses. Methods Our application consists of three supervised learning models: (1) nerve fascicle instance segmentation, (2) myelinated fibre detection and (3) myelin sheath segmentation. We fine‐tuned these models using 86 toluidine blue‐stained slides from various neuropathies and developed an open‐source Python library. Results Performance evaluation showed (1) a mask average precision (AP) of 0.861 for fascicle segmentation, (2) box AP of 0.711 for fibre detection and (3) a mean intersection over union (mIoU) of 0.817 for myelin segmentation. Our software identified 323,298 nerve fibres and 782 fascicles in 70 WSIs. Small and large fibre populations were objectively determined based on clustering analysis. The demyelination group had large fibres with thinner myelin sheaths and higher g‐ratios than the vasculitis group. The slope of the regression line from the scatter plots of the diameters and g‐ratios was higher in the demyelination group than in the vasculitis group. Conclusion We developed an application that performs whole slide morphometry of human biopsy samples. Our open‐source software can be used by clinicians and pathologists without specific machine learning skills, which we expect will facilitate data‐driven analysis of sural nerve biopsies for a more detailed understanding of these diseases.