Estimation of myocardial and blood gadolinium concentrations from T1 mapping via pharmacokinetic modeling: influence of elastic deformation registration

Abstract Accurate quantification of myocardial gadolinium-based contrast medium (CM) concentration is important for longitudinal cardiac MRI assessment. We developed a T1 mapping-based pharmacokinetic (PK) approach incorporating elastic deformation registration (ER) to improve fitting accuracy. Forty-nine participants (median age, 62 years) underwent native and dynamic T1 mapping at 2, 5, 9, and 15 minutes after CM administration. CM concentration maps for myocardium and blood pool were generated at each time point, with and without ER, and fitted to a PK model. Model performance was evaluated using the coefficient of determination (R²), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Agreement between measured myocardial CM concentrations and model estimates was assessed, and the mean difference between them was calculated. T1-derived CM concentrations showed excellent agreement with phantom reference values (R 2 = 0.999). ER improved model fit compared with no registration, with smaller absolute mean differences at all time points: 0.026 vs. 0.029 mM/l (2 min), –0.021 vs. –0.025 (5 min), –0.022 vs. –0.023 (9 min), and 0.018 vs. 0.021 mM/l (15 min). This ER-based dynamic T1 mapping approach improved the accuracy of PK modeling for myocardial CM concentration, supporting its potential utility for refined myocardial tissue characterization.