MRI-guided prostate radiation therapy planning: Investigation of dosimetric accuracy of MRI-based dose planning

Background and purpose Dose planning requires a CT scan which provides the electron density distribution for dose calculation. MR provides superior soft tissue contrast compared to CT and the use of MR-alone for prostate planning would provide further benefits such as lower cost to the patient. This study compares the accuracy of MR-alone based dose calculations with bulk electron density assignment to CT-based dose calculations for prostate radiotherapy. Materials and methods CT and whole pelvis MR images were contoured for 39 prostate patients. Plans with uniform density and plans with bulk density values assigned to bone and tissue were compared to the patient’s gold standard full density CT plan. The optimal bulk density for bone was calculated using effective depth measurements. The plans were evaluated using ICRU point doses, dose volume histograms, and Chi comparisons. Differences in spatial uniformity were investigated for the CT and MR scans. Results The calculated dose for CT bulk bone and tissue density plans was 0.1 ± 0.6% (mean ± 1 SD) higher than the corresponding full density CT plan. MR bulk bone and tissue density plans were 1.3 ± 0.8% lower than the full density CT plan. CT uniform density plans and MR uniform density plans were 1.4 ± 0.9% and 2.6 ± 0.9% lower, respectively. Paired t -tests performed on specific points on the DVH graphs showed that points on DVHs for all bulk electron density plans were equivalent with two exceptions. There was no significant difference between doses calculated on Pinnacle and Eclipse. The dose distributions of six patients produced Chi values outside the acceptable range of values when MR-based plans were compared to the full density plan. Conclusions MR-alone bulk density planning is feasible provided bone is assigned a density, however, manual segmentation of bone on MR images will have to be replaced with automatic methods. The major dose differences for MR bulk density plans are due to differences in patient external contours introduced by the MR couch-top and pelvic coil.