Quantitative AI Volumetry Profiling Regional Muscle Hypotrophy in Chronic Limb-Threatening Ischemia

Chronic limb-threatening ischemia (CLTI) leads to severe functional impairment and degeneration of lower-limb musculature. Although muscle wasting in CLTI has been assessed using single-slice or density-based approaches, three-dimensional volumetric analysis using artificial intelligence (AI) remains limited. This study aimed to quantify and compare muscle volumes in both limbs of patients with unilateral CLTI using AI-based computed tomography angiography segmentation to identify muscle groups most susceptible to ischemic hypotrophy.

The cohort of 100 consecutive patients undergoing lower-limb computed tomography angiography for CLTI evaluation was screened. After exclusion of bilateral disease, prior amputation, poor image quality, and nonocclusive pathology, 30 patients with unilateral CLTI (Fontaine III-IV) and 10 controls without significant stenosis were included. Three-dimensional muscle volumes were obtained using a deep-learning model (TotalSegmentator 2.11.0). The hypotrophy rate was defined as the ischemic-to-contralateral limb volume ratio. Significant hypotrophy was determined using a control-derived threshold of 2 standard deviations. Sarcopenia was evaluated using psoas muscle area.

Significant muscle hypotrophy was present in 11 of 30 CLTI patients (37%). The greatest volume loss involved the gluteus minimus, quadriceps femoris, medial thigh compartment, and sartorius (P < 0.05). No significant hypotrophy was observed in the gluteus maximus, gluteus medius, or lower-leg muscles. Proximal muscle volume loss correlated strongly with total limb hypotrophy, particularly in the quadriceps (ρ = 0.90) and gluteal region (r = 0.97). Sarcopenia increased the odds of significant hypotrophy more than 10-fold (odds ratio 10.29; P = 0.047).

AI-based volumetry enables precise, operator-independent quantification of regional muscle wasting in CLTI.