Cloud-Enabled Automated CT Assessment of Pelvic Muscle Quality in Women With and Without Low-Energy Femoral Neck Fracture.

This retrospective study aimed to compare pelvic muscle parameters between female patients with and without femoral neck fractures and to evaluate the feasibility of an automated CT-based workflow for exploring associations between pelvic muscle quality measures and femoral neck fracture status. The study included 119 female patients with low-energy femoral neck fractures and 107 age-matched female controls. Non-contrast computed tomography images were analyzed using a cloud-integrated deep learning segmentation tool. Pelvic muscle volume and quality parameters, including intramuscular adipose tissue, myosteatosis, and functional lean muscle, were quantified for the iliopsoas and gluteal muscles. Age-adjusted multivariate analysis, receiver operating characteristic analysis, and logistic regression were performed. Age and cortical bone parameters did not differ significantly between groups. In contrast, pelvic muscle quality parameters showed significant between-group differences after adjustment for age. Iliopsoas myosteatosis demonstrated the highest individual discriminatory performance, with an area under the curve of 0.711. The combined four-muscle myosteatosis model achieved the highest apparent within-sample discrimination, with an area under the curve of 0.739, indicating moderate discriminatory performance. Iliopsoas MYO values above the ROC-derived cut-off were associated with higher odds of belonging to the fracture group (OR, 4.22). Automated measurements showed excellent reliability. These findings suggest that pelvic muscle quality measures may provide complementary information for fracture-related assessment on routine CT examinations. Automated CT-based muscle quality assessment may provide a reproducible approach for opportunistic body-composition analysis on routine CT examinations. The cloud-enabled workflow used in this study illustrates technical feasibility; however, its clinical utility for fracture-risk stratification requires prospective validation.