Click here to


Are you sure ?

Yes, do it No, cancel

Can 0.345T Magnetic Resonance Images Be Used in Breast Cancer for Post Lumpectomy Target Delineation?

P Yadav1*, H Musunuru1 , J Witt1 , A Kuczmarska-Haas1 , B Anderson1 , F Kelcz2 , P Mahler1 , (1) Department of Human Oncology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, (2) Department of Radiology, School of Medicine and Public Health, University of Wisconsin, Madison, WI


(Monday, 7/15/2019) 3:45 PM - 4:15 PM

Room: Exhibit Hall | Forum 5

Purpose: Feasibility of utilizing 0.345T MR scans in breast cancer for post lumpectomy target delineation.

Methods: Fifteen breast cancer patients treated on ViewRay MRIdian Linac were selected for this study. An experienced breast radiologist (BR) and a breast radiation oncologist (RO) were asked to designate individual cavity visualization scores (CVS) for CT-TB and MR-TB by utilizing CT contours as reference contours. For this study CT contours were designated as reference contours. Contour volumes and conformity indices (CI) including Dice coefficient (DC), Geographical Miss Index (GMI) and Discordant index (DI) were calculated. Chi-squared test and Wilcoxon signed-rank test were used to assess the difference between CT and MR contours.

Results: Proportion of patients having adequate CVS (3-5) was higher for CT dataset when compared to MR dataset (73% vs. 47%) for both BR and RO. Median TB volume was slightly lower in the MR contours for BR data (MR 13.8cc (IQR5.8-22.9) vs. CT 15.5cc (10.6-34.0)), whereas it was significantly lower in the CT contours for RO data (MR 108.2cc (30.3-178.4) vs. CT 66.6cc(26.2-113.4)). Median DC was consistently higher for RO contours (BR 0.38 vs. RO 0.5). Median GMI was higher for BR dataset vs. RO contours (0.72(0.49-0.99) vs. 0.4(0.13-0.56)), indicating that in the BR dataset, MR contours consistently did not include a significant proportion of CT contours. Median DI was 0.48 and 0.54 for BR and RO contours, indicating that MR contours had significant amount of information not present in the CT contours.

Conclusion: CVS and TB volumes were lower for MR dataset in comparison to the CT dataset and shows consistency with diagnostic MR scans. In the era of MR-guided adaptive radiotherapy, a prospective study integrating 3T MR, 0.35T MR and CT might provide information about the optimal imaging modality for treatment bed delineation for breast cancer patients.


Breast, Radiation Therapy, Target Localization


Not Applicable / None Entered.

Contact Email