European Journal of Radiology · Apr 15, 2026

Longitudinal trajectory of thoracic CTA-derived sarcopenia stratifies mortality risk in transcatheter aortic valve replacement

By Zhang Y, Zheng T, Xu Y et al.

BACKGROUND: Although sarcopenia is a known predictor of outcomes in cardiovascular diseases, its disease trajectory, including improvement, stability, or worsening, following transcatheter aortic valve replacement (TAVR) remains insufficiently explored. This study pioneered using pectoralis muscle index (PMI) from routine thoracic computed tomography angiography (CTA) to track sarcopenia changes post-TAVR and determine prognostic value. METHODS: This single-center cohort included consecutive TAVR patients with serial thoracic CTA at baseline and 1 year. PMI was calculated as pectoralis muscle area at the fourth thoracic vertebra normalized to height². Using sex-specific PMI thresholds, patients were stratified into four phenotypes: non-sarcopenic, early-onset, late-onset, or persistent sarcopenia. Multivariate Cox regression identified mortality predictors. RESULTS: Among 258 patients (median age 72.86; 56.59% male), both baseline PMI (median 1,067.19 mm²/m², p = 0.038) and 1-year PMI (median 1,145.95 mm²/m², p = 0.016) independently predicted survival during a median 1,495-day follow-up (mortality 12.79%). Compared to non-sarcopenic patients, those with persistent sarcopenia (17.83% of cohort) had significantly elevated mortality risk (Hazard ratio = 3.94, 95% CI = 1.44-10.80, p = 0.008). CONCLUSIONS: This study suggests the potential of PMI quantification via thoracic CTA as a clinically integrable tool for prognostic stratification in TAVR recipients through sarcopenia monitoring, which warrants further randomized investigation for potential generalization to other cardiac interventions. The identification of persistent sarcopenia as a modifiable risk factor mandates the implementation of protocolized nutritional optimization and structured rehabilitation programs in high-risk subgroups.

doi.org/10.1016/j.ejrad.2026.112873 PubMed 42008889 Biomarkers