MyoSOTHES

MyoSOTHES performs automated segmentation and quantitative analysis of skeletal muscle fibers in Hematoxylin-Eosin (HE) stained histological sections to enable measurement of fiber size distributions and identification of centrally nucleated fibers.

Key Features:

• Integration of Cellpose and QuPath: Integrates Cellpose and QuPath into a workflow tuned for HE-stained sections to perform segmentation and downstream analysis.
• Improved segmentation consistency: Addresses Cellpose default limitations for large size variations among fibers, increasing detection consistency and accuracy.
• Quantitative performance gains: Reports an F1-score improvement from 0.801 to 0.919 and a 31% reduction in RMSE on muscle fiber diameter measurements.
• Muscle health metrics: Quantifies Feret's diameter distribution and identifies Centrally Nucleated Fibers (CNF) for assessment of muscle condition and treatment effects.
• Validation in disease model: Validated in an animal gene transfer study in a model of -Sarcoglycanopathy, demonstrating dose-response detection consistent with published findings and suitability for retrospective analysis of archival HE sections.

Scientific Applications:

• Muscular dystrophy and myopathy research: Quantitative assessment of fiber size distributions and CNF in studies of muscular dystrophies and other myopathies.
• Treatment and gene therapy evaluation: Detection and quantification of dose-response effects in gene transfer and related therapeutic studies.
• Retrospective histology analysis: Reanalysis of archival HE-stained muscle sections to extract quantitative metrics from historical datasets.

Methodology:

Combines Cellpose deep-learning segmentation with QuPath-based processing in a workflow tuned for HE staining; performance evaluated using F1-score and RMSE on fiber diameter.