What question did this study set out to answer?

The aim is to develop a model for quantifying muscle volumes to predict functional recovery after hip arthroplasty.

June 20, 2026Scientific ReportsOpen Access

A two-stage deep-learning model using MobileNetV2 and U-Net for CT-based muscle volume assessment in total hip arthroplasty

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial developed a two-stage deep-learning model to assess muscle volumes, implying improved functional recovery insights in THA.

The aim is to develop a model for quantifying muscle volumes to predict functional recovery after hip arthroplasty.
Developed a two-stage deep-learning framework combining slice-level classification and U-Net segmentation.
Trained using CT scans from 107 patients undergoing primary total hip arthroplasty.
Validated against manual segmentation and applied to an independent cohort of 58 patients.
Automated segmentation showed high agreement with manual measurements (absolute volume errors of 1-10%, Dice coefficients > 0.89).
Preoperative muscle volumes were significantly lower on the affected side compared to the unaffected side.
Greater volumes of gluteus maximus, quadriceps, and hamstrings were significantly associated with improved postoperative strength and performance.

Tachibana et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

Randomized trial developed a two-stage deep-learning model to assess muscle volumes, implying improved functional recovery insights in THA.