Feasibility and accuracy of quantitative imaging on a 1.5 T MR-linear accelerator

Purpose: Systems for magnetic resonance (MR-) guided radiotherapy enable daily MR imaging of cancer patients during treatment, which is of interest for treatment response monitoring and biomarker discovery using quantitative MRI (qMRI). Here, the performance of a 1.5 T MR-linac regarding qMRI was assessed on phantoms. Additionally, we show the feasibility of qMRI in a prostate cancer patient on this system for the first time. Materials and methods: Four 1.5 T MR-linac systems from four institutes were included in this study. T ₁ and T ₂ relaxation times, and apparent diffusion coefficient (ADC) maps, as well as dynamic contrast enhanced (DCE) images were acquired. Bland–Altman statistics were used, and accuracy, repeatability, and reproducibility were determined. Results: Median accuracy for T ₁ ranged over the four systems from 2.7 to 14.3%, for T ₂ from 10.4 to 14.1%, and for ADC from 1.9 to 2.7%. For DCE images, the accuracy ranged from 12.8 to 35.8% for a gadolinium concentration of 0.5 mM and deteriorated for higher concentrations. Median short-term repeatability for T ₁ ranged from 0.6 to 5.1%, for T ₂ from 0.4 to 1.2%, and for ADC from 1.3 to 2.2%. DCE acquisitions showed a coefficient of variation of 0.1–0.6% in the signal intensity. Long-term repeatability was 1.8% for T ₁ , 1.4% for T ₂ , 1.7% for ADC, and 17.9% for DCE. Reproducibility was 11.2% for T ₁ , 2.9% for T ₂ , 2.2% for ADC, and 18.4% for DCE. Conclusion: These results indicate that qMRI on the Unity MR-linac is feasible, accurate, and repeatable which is promising for treatment response monitoring and treatment plan adaptation based on daily qMRI.