Click here to


Are you sure ?

Yes, do it No, cancel

Initial Performance Results of a Commercial MR-Only Workflow for Pelvis

C Gage1 , T Scheer2 , M Requardt3 , E Paulson4*, (1) Medical College of Wisconsin, Milwaukee, WI, (2) Siemens Healthineers, Erlangen, Germany, (3) Siemens Healthineers, Erlangen, Germany, (4) Medical College of Wisconsin, Milwaukee, WI


(Sunday, 7/29/2018) 4:00 PM - 4:55 PM

Room: Karl Dean Ballroom B1

Purpose: We demonstrate here initial performance results of a commercial MR-only workflow for pelvis.

Methods: Following CT simulation, four patients (one prostate, three prostate bed) were imaged on a Siemens Aera using a clinical MR simulation protocol including a 3D Dixon sequence. An atlas of fully contoured, standardized, water-only Dixon images was constructed in MIM (MIM Software) for MR-based auto-contouring. Synthetic CTs were generated by segmenting in-/opposed-phase and fat-/water-only Dixon images using a thresholding/bone atlas algorithm (Siemens). The synthetic CTs were transferred to Monaco (Elekta) where original VMAT plans were templated and recalculated on the synthetic CTs. To control for dose differences due to anatomical variation, original VMAT plans were also templated and recalculated on daily kV fan-beam CTs obtained from an in-room Siemens Definition CT-on-rails (CTOR). Dose distributions were compared using 3D gamma and DTA analysis (3%/3mm/5% threshold) in 3DVH (Sun Nuclear). D95PTV and OAR dose-volume constraints were also compared.

Results: Overall segmentation accuracy of synthetic CT images was good, with the thresholding/bone atlas algorithm correctly identifying air in the rectum when appropriate. D95PTV percent differences were within 1.0%. 3D dose distribution matching rates exceeded 95% for both gamma and DTA analysis on the synthetic CTs. In comparison, matching rates ranged from 71% - 92% for plans recalculated on daily CTOR images. These results demonstrate the challenge of validating synthetic CTs in the pelvis and suggest that anatomical variations between synthetic and planning CTs may dominate dose differences compared to minor segmentation or density assignment errors. DRRs generated from the synthetic CTs were successfully transferred to Mosaiq (Elekta). Synthetic CTs were successfully loaded as IGRT reference images in XVI (Elekta).

Conclusion: Preliminary validation of a commercial MR-only workflow for pelvis revealed good feasibility and performance. Further validation using the reversed simulation workflow will be the focus of future work.

Funding Support, Disclosures, and Conflict of Interest: Partially supported by Siemens Healthineers.


MRI, Treatment Planning, Prostate Therapy


IM/TH- MRI in Radiation Therapy: MRI for treatment planning

Contact Email