Vascular-augmented two-compartment fitting improves model performance for intermittent myocardial T1 mapping

Conventional gadolinium-enhanced cardiac MRI typically evaluates myocardial tissues at a single post-contrast time point. In contrast, dynamic T1 mapping enables the estimation of contrast agent concentrations and subsequent pharmacokinetic modeling. This study compared a normal composite two-compartment model incorporating myocardial vascular components with the conventional Brix model.

This retrospective study included 107 participants who underwent dynamic T1 mapping at 4 points after contrast administration. Contrast agent concentrations derived from T1 maps were fitted using the Brix and composite pharmacokinetic models. Model performance was assessed using the residual sum of squares (RSS), Akaike information criterion (AIC), and Bayesian information criterion (BIC), along with spatial comparison of model-estimated concentration maps.

The composite model exhibited significantly lower RSS, AIC, and BIC values than the Brix model (all p < 0.001). Absolute parameter estimation errors were reduced across all time points. In addition, systematic spatial differences in estimated myocardial contrast concentrations were observed between the two models, indicating distinct model-dependent representations of longitudinal contrast kinetics.

The composite model achieved superior fitting performance compared with the Brix model. Explicit incorporation of vascular kinetics improves the longitudinal characterization of contrast behavior and enhances quantitative assessment of myocardial tissue properties.