2840 - Assessing Clinical Viability of Contour Based Hybrid Deformable Registration (CBH DIR) through Analysis of Spatial Dose Mapping of Per Structure High Dose Regions and Contour Transfer Fidelity for GYN HDR/EBRT Combined Treatment

Purpose/Objective(s): The traditional method of summing max doses to evaluate the risk of toxicity from HDR/EBRT treatments is conservative as these max doses may not overlap. HDR applicators deform nearby OARs, and rigid fusion (RF) does not account for those changes. A deformable registration-based (DIR) method accounts for changes in anatomy, allowing for the mapping of dose between images to see if high dose regions overlap. This study investigates using the spatial mapping of max doses and contour transfer accuracy as tools to QA registrations for GYN HDR/EBRT dose sum.

Materials/Methods: A CBH DIR, using bladder, rectum, and sigmoid contours to guide registration of key OARs, was tested. Five HDR/EBRT CT datasets with various applicators were tested by four medical physicists from three clinics. Qualitative accuracy of max dose and D2cc contour mapping (1-5 scale) was assessed, along with overall clinical viability. Dice Similarity Coefficients (DSC) compared EBRT contours transferred onto the HDR CT to the native HDR contours for both RF and DIR. Testers evaluated whether the tools aided in determining DIR viability.

Results: The DIR was rated as clinically viable by all testers in test cases 1 and 2, and by 4 out of 5 testers in test cases 3-5. All testers agreed that the tools provided the info needed to determine the viability of the DIR for dose sum.

Conclusion: Providing a way to evaluate DIR and its impact on dose sum is often as valuable as the accuracy of the registration itself because spatial dose data informs decision making in the treatment process. Our results indicate that dosimetric tools enable confidence in the registration in key areas by allowing the user to confirm the transferred max dose or D2cc represents the same anatomical location in both images. The results also show the CBH DIR is promising, demonstrating improved contour transfer accuracy over RF, and positive qualitative accuracy ratings for max dose and D2cc. This is encouraging, given the challenging deformations of OARs due to presence/absence of the HDR applicator between images. Overall, this study provides a promising framework for evaluating DIR for spatial dose sum for GYN HDR/EBRT treatments. Future work should include a larger sample size of testers/datasets. Additionally, further DIR research and development, including quantitative analysis of registration error and multimodality support for MR/CT image pairs, could improve overall robustness of this registration-based dose sum method.