ESTRO 2025 - Abstract Book

S3001

Physics - Image acquisition and processing

ESTRO 2025

Results: The time-efficient DESPOT1 and MIRACLE methods showed relatively large bias and precision (Figure 1, top row, T1 spheres only), suggesting model limitations for samples with T2 > 500 ms. Better performance was obtained for vials with more physiologically relevant T1 and T2 values, T2 < 500 ms and T1 > 100 ms (Figure 1, bottom row, selected T1 and T2 spheres).

For vial 5 (T1=514; T2=446 ms), T1 bias was 42, -57, -6, and 2 ms for MIRACLE, DESPOT1, MIX and IR respectively. Repeatability and reproducibility were 61 ms and 12% for MIRACLE, and 13 ms and 3% for DESPOT1. Results compare favorably to QIBs in literature: reproducibility is 17% for ADC, 20% for k-trans in prostate [9]. Conclusion: Consensus protocols have been developed for T1 relaxometry on the Unity MR-Linac. Both the DESPOT1 and MIRACLE techniques show promise for T1 mapping in vivo, based on relative time efficiency and adequate test performance using the NIST phantom.

Keywords: functional imaging, MR-Linac, multi-institution

References: [ 1 ] Raaymakers et al. Phys Med Biol 2017 62 , L41–L50. [ 2 ] Bentzen et al. Semin Radiat Oncol 2011 21 , 101–110. [ 3 ] Russek et al. Proc Int Soc Magn Reson Med 2012 20 , 2456.