High-Resolution, Free-Running, Whole Heart Fat Fraction Mapping to Monitor Fatty Remodeling and Therapeutic Response to Semaglutide in a Porcine HFpEF Model

Background: Heart failure with preserved ejection fraction (HFpEF) is increasingly linked to myocardial lipid accumulation, contributing to stiffness, impaired relaxation, and adverse remodeling. GLP-1 receptor agonists (GLP-1RA) show clinical benefits in HFpEF, including epicardial adipose tissue (EAT) reduction¹, but their effects on intramyocardial fat remain unknown. Because intramyocardial fat is heterogeneously distributed, spatially resolved, noninvasive techniques to monitor lipid burden and therapeutic modulation are limited. Multi-echo gradient-echo (mGRE) enables water–fat separation, but conventional methods are hindered by resolution, breath-hold demands, and off-resonance artifacts. We propose a free-running, high-resolution whole-heart mGRE approach for fat-fraction mapping in a porcine HFpEF model to evaluate semaglutide’s therapeutic effects.

Methods: An established “Two-Hit” HFpEF Gottingen Minipig model² (N=8) was studied, with echocardiography confirming phenotype. Four animals received semaglutide (4 nmol/kg, IM, weekly). CMR was performed on a 3T system (MAGNETOM Biograph mMR, Siemens) at baseline and 3 months. A free-breathing, non-ECG-gated whole-heart mGRE sequence (8 min, 1.6 mm isotropic)^3,4 was prescribed with conventional cine. mGRE images were processed with IDEAL-constrained estimation to generate fat-fraction maps⁵. Echocardiography assessed diastolic function. LVEF, E/e’ ratio, intramyocardial FF (imFF), and EAT volume were quantified, normalized to baseline, and compared between groups.

Results: All animals successfully completed baseline and 3-month follow-up imaging without complications. Cine CMR confirmed preserved systolic function, with no significant change in LVEF between baseline and follow-up (57.9±9.7% vs. 55.1±14.4%, P=0.35). At baseline, imFF and EAT volumes were comparable between treated and untreated groups (imFF: 5.8±1.6% vs. 5.5±0.9%, P=0.745; EAT: 34.3±20.5 mL vs. 36.1±2.9 mL, P=0.869). By 3 months, untreated animals exhibited significant increases in both imFF (1.94±0.79%, P=0.016) and EAT volume (13.0±4.6 mL, P=0.011), consistent with progressive fatty remodeling. In contrast, semaglutide-treated pigs demonstrated stable imFF and EAT expansion (imFF:Δ0.40±0.76% vs.Δ1.94±0.79%, P=0.032; EAT: Δ3.4±5.3 mL vs.Δ13.0±4.6 mL, P=0.019). Echocardigraphy findings confirm the semaglutide treatment in improving diastolic function (E/e’ ratio:Δ-3.34±0.79 vs.Δ3.47±0.93, P< 0.001). These findings indicate that semaglutide therapy mitigated both intramyocardial and epicardial fat accumulation compared with untreated animals.

Conclusion: We demonstrate the feasibility of free-running whole-heart fat-fraction mapping within 8 minutes, enabling spatially resolved assessment of myocardial lipid remodeling. This is the first report of noninvasive imaging capturing reduced intramyocardial fat accumulation with semaglutide therapy in HFpEF, highlighting translational potential for monitoring myocardial lipid burden and guiding metabolic interventions.

Figure 1. Schematic overview of the HFpEF porcine model, study design, and imaging workflow. Animals with “Two-Hit” induced HFpEF underwent baseline and 3-month follow-up CMR. Pigs were randomized into untreated and semaglutide-treated groups. Cine CMR and the proposed free-running mGRE sequence were performed to evaluate cardiac function and quantify myocardial and epicardial fat composition, enabling assessment of therapeutic effects on functional remodeling and lipid burden. Echocardiography confirming diastolic heart function of disease phenotype.
Figure 2. Quantitative measurements of ejection fraction (EF), early mitral flow velocity/early diastolic mitral annular velocity (E/e’ ratio), intramyocardial fat fraction (imFF), and epicardial adipose tissue (EAT) volume in HFpEF animals with and without semaglutide therapy. While EF remained unchanged between groups, semaglutide-treated animals demonstrated significantly attenuated increases in both imFF and EAT compared with untreated animals. These findings highlight the ability of the proposed whole-heart fat-fraction mapping to detect therapeutic effects not evident on conventional cine imaging (*P < 0.05, **P < 0.01, ***P < 0.001).