Residual neural network simultaneous cardiac T1⍴ and T2 mapping using a combined spin echo, spin lock pulse sequence

Motivation: Accurate cardiac T1ρ and T2 mapping is crucial for diagnosing myocardial conditions. Current methods are time-consuming and limited. Goal(s): Develop an accurate, efficient method for simultaneous and accurate T1rho and T2 mapping in cardiac MRI using a single acquisition. Approach: A combined spin-echo/spin-lock pulse sequence with a residual neural network trained on segmented myocardium data to simultaneously estimate T1rho and T2 values. Results: The proposed method is able to accurately estimate T1rho and T2 maps, with values that closely match reference measurements, and reduces estimation bias and variability in T1rho and T2 values compared to traditional methods. Impact: This method enhances cardiac imaging by enabling simultaneous T1rho and T2 mapping in a single acquisition, reducing scan time and improving accuracy. It provides more consistent myocardial tissue characterization, potentially aiding early diagnosis and treatment planning for heart disease.