Modified Look-Locker T1 evaluation using Bloch simulations: human and phantom validation.

Modified Look-Locker imaging is frequently used for T(1) mapping of the myocardium. However, the specific effect of various MRI parameters (e.g., encoding scheme, modifications of flip angle, heart rate, T(2), and inversion times) on the accuracy of T(1) measurement has not been studied through Bloch simulations. In this work, modified Look-Locker imaging was characterized through a numerical solution for Bloch equations. MRI sequence parameters that may affect T(1) accuracy were systematically varied in the simulation. For validation, phantoms were constructed with various T(2) and T(1) times and compared with Bloch equation simulations. Human volunteers were also evaluated with various pulse sequences parameters to assess the validity of the numerical simulations. There was close agreement between simulated T(1) times and T(1) times measured in phantoms and volunteers. Lower T(2) times (i.e., <30 ms) resulted in errors greater than 5% for T(1) determination. Increasing maximum inversion time value improved T(1) accuracy particularly for precontrast myocardial T(1). Balanced steady-state free precession k space centric encoding improved accuracy for short T(1) times (post gadolinium), but linear encoding provided improved accuracy for precontrast T(1) values. Lower flip angles are preferred if the signal-to-noise ratio is sufficiently high. Bloch simulations for modified Look-Locker imaging provide an accurate method to comprehensively quantify the effect of pulse sequence parameters on T(1) accuracy. As an alternative to otherwise lengthy phantom studies or human studies, such simulations may be useful to optimize the modified Look-Locker imaging sequence and compare differences in T(1)-derived measurements from different scanners or institutions.