366 - MR Multitasking-Driven Abdominal Integrated Imaging (MT-AI²): Early Clinical Experience and Variability of Quantitative Maps

Purpose/Objective(s): Liver function assessment is vital for guiding SBRT planning to prevent functional liver damage and radiation-induced liver disease. Traditional methods lack regional specificity, while MR T1/T2 mapping with liver-specific contrast offers a promising alternative. Our group developed a novel MT-AI² sequence, a multi-contrast-weighted 4D-MR that provides T1 and T2 maps in one 8-min scan. This study reports early clinical experience and T1/T2 variations compared to published MR sequences. This is a baseline study or future multi-site trials adopting different MR T1/T2 mapping sequences.

Materials/Methods: Six subjects (4 patients and 2 healthy volunteers) were recruited for intra- and/or inter-session MT-AI² repeatability study. All 6 subjects underwent MT-AI² twice in the same MR session and the 2 healthy volunteers also received MT-AI² on different days. Additional 45 liver cancer patients underwent same-day CT and three free-breathing MR mapping scans on a 3T MR simulator: MT-AI² (T1/T2 maps), MP-GRASP (T1), and RadTSE (T2). Twenty-one were imaged pre- and 24 were imaged at least 15 mins post-Eovist™ injection. Clinical GTV was transferred to the MR maps by rigid registration. T1 and T2 values within GTV and liver minus (-) GTV were extracted. In patients with prior liver-directed therapies, only the current GTV was analyzed. Five random points within liver-GTV were measured, excluding blood vessels and peripheral liver regions. Calibration curves between the different mapping sequences were developed using regression analysis.

Results: All 6 subjects showed excellent intra-session reproducibility (T1 % difference mean ± std: 3.8% ± 2.5% difference; T2: 8.2% ± 4.9%). The inter-session variability was more pronounced (T1 % difference mean ± std: 11.9% ± 8.9%; T2: 16.2% ± 19.1%). Strong correlations were noted for T1, particularly in GTVs. Post-contrast GTV T1 had the highest correlation (R² = 0.87, RMSE = 69.41 ms), followed by pre-contrast GTV T1 (R² = 0.72, RMSE = 112.88 ms) and Liver-GTV (pre-contrast R² = 0.50, RMSE = 86.38 ms; post-contrast R² = 0.39, RMSE = 39.63 ms). T2 mapping displayed greater variability overall. Pre-contrast GTV T2 had a moderate correlation (R² = 0.63, RMSE = 6.52 ms), while post-contrast GTV T2 correlation was lower (R² = 0.40, RMSE = 10.01 ms). For liver-GTV, T2 values showed slightly better consistency, with pre-contrast at R² = 0.73 (RMSE = 2.63 ms) and post-contrast at R² = 0.33 (RMSE = 5.61 ms). The difference in contrast time-lapse across sequences performed in each patient may have contributed to the weaker correlation of post-contrast T1/T2 values in liver-GTV.

Conclusion: We implemented and validated novel MT-AI² for simultaneous liver T1/T2 mapping. T1 mapping using MT-AI² demonstrated a robust linear correlation with the published method. Conversely, T2 mapping displayed more variability, especially following contrast administration, necessitating additional caution when converting between the published method and MT-AI².