Distinct Biomarker Signatures of CMR-Defined Myocardial Remodeling in Moderate Aortic Stenosis

Background: Moderate aortic stenosis (AS) is associated with adverse outcomes, yet the myocardial response to pressure overload varies substantially between patients and is incompletely reflected by valve hemodynamics. We aimed to determine whether circulating biomarkers of myocardial stress, injury, and fibrosis associate with cardiac magnetic resonance (CMR)-defined myocardial remodeling in moderate AS. Methods: Eighty-nine patients with moderate AS prospectively underwent CMR (cine imaging, T1 mapping, extracellular volume fraction [ECV], late gadolinium enhancement [LGE]) with concurrent measurement of NT-proBNP, high-sensitivity troponin T, sST2, TIMP-1, and Galectin-3. Biomarker associations with the phenotype-defining CMR parameters were tested with adjustment for age, sex, and renal function and correction for multiple testing, and patients were additionally classified into 4 hierarchical remodeling phenotypes. Results: Six of 20 biomarker-CMR associations remained significant after correction for multiple testing. NT-proBNP tracked LV mass index (beta=0.50), ECV (beta=0.35), and LGE burden (beta=0.43, all q<0.05). TIMP-1 and Galectin-3 were selectively associated with native T1 (both beta=0.35, q=0.012), and sST2 was inversely associated with LV mass index (beta=-0.27, q=0.046). TIMP-1 (adjusted OR 2.18 per SD, 95% CI 1.37-3.48, q=0.005) and NT-proBNP (OR 2.10, 1.19-3.69, q=0.026) increased with higher hierarchical remodeling phenotype. No biomarker was associated with echocardiographic mean gradient, peak velocity, or indexed valve area after correction. Conclusions: Circulating biomarkers showed distinct associations with CMR measures of myocardial remodeling in moderate AS. Matrix-related biomarkers were associated with native T1, whereas NT-proBNP was associated across several remodeling domains. These signals were not measurably related to valve hemodynamics.