Encrypted login | home

Program Information

MRI-Based Radiotherapy Treatment Planning In Pelvis

no image available
S Hsu

S Hsu*, Y Cao , S Jolly , J Balter , University of Michigan, Ann Arbor, MI

Presentations

SU-C-17A-1 Sunday 1:00PM - 1:55PM Room: 17A

Purpose: To support radiotherapy dose calculation, synthetic CT (MRCT) image volumes need to represent the electron density of tissues with sufficient accuracy. This study compares CT and MRCT for pelvic radiotherapy.
Methods: CT and multi-contrast MRI acquired using T1-based Dixon, T2 TSE, and PETRA sequences were acquired on an IRB-approved protocol patient. A previously published method was used to create a MRCT image volume by applying fuzzy classification on T1-weighted and calculated water image volumes (air and fluid voxels were excluded using thresholds applied to PETRA and T2-weighted images). The correlation of pelvic bone intensity between CT and MRCT was investigated. Two treatment plans, based on CT and MRCT, were performed to mimic treatment for: (a) pelvic bone metastasis with a 16MV parallel beam arrangement, and (b) gynecological cancer with 6MV volumetric modulated arc therapy (VMAT) using two full arcs. The CT-calculated fluence maps were used to recalculate doses using the MRCT-derived density grid. The dose-volume histograms and dose distributions were compared.
Results: Bone intensities in the MRCT volume correlated linearly with CT intensities up to 800 HU (containing 96% of the bone volume), and then decreased with CT intensity increase (4% volume). There was no significant difference in dose distributions between CT- and MRCT- based plans, except for the rectum and bladder, for which the V45 differed by 15% and 9%, respectively. These differences may be attributed to normal and visualized organ movement and volume variations between CT and MR scans.
Conclusion: While MRCT had lower bone intensity in highly-dense bone, this did not cause significant dose deviations from CT due to its small percentage of volume. These results indicate that treatment planning using MRCT could generate comparable dose distributions to that using CT, and further demonstrate the feasibility of using MRI-alone to support Radiation Oncology workflow.

Funding Support, Disclosures, and Conflict of Interest: NIH R01EB016079


Contact Email: