Program Information
Online Adaptive Replanning Based On Voxel-By-Voxel Prescription Map
E Ahunbay*, X Chen , X Li , Medical College of Wisconsin, Milwaukee, WI
Presentations
MO-RPM-GePD-J(A)-2 (Monday, July 31, 2017) 3:45 PM - 4:15 PM Room: Joint Imaging-Therapy ePoster Lounge - A
Purpose: With the introduction of high-field MR-Linac, spatial tumor and response heterogeneity can be obtained during the radiation therapy (RT) delivery. This work aims to develop an online replanning strategy to account for the changing spatial heterogeneity.
Methods: To demonstrate the idea, realistic MRI/ADC maps acquired at different time points during pancreatic cancer RT were converted to the required, voxel-based dose map. A reference plan was generated to achieve uniform TCP over all voxels in GTV using a research planning system (Monaco, Elekta) based on the planning ADC map. The reference plan was used as in input for fast replanning based on daily MRI sets using Warm Start Optimization (WSO) algorithm inplementd in Monaco. The doses in voxels outside of the target were obtained by deformably converting the reference doses. An in-house software was developed to iteratively perform WSO by updating the goal dose distribution after each iteration to maximize the combined tumor TCP (product of all voxels’ TCP values). Non-uniform TCP was allowed. At each iteration of WSO, the goal doses for the voxels with the low TCPs were increased, resulting in increased doses to these voxels by WSO's next run. Resulting doses were rescaled to keep volumes enclosed by 5-50Gy isodose lines in the non-tumor region same or smaller than those in the reference plan, and iteration stops when no more increase in TCP is achievable.
Results: The method was successfully implemented. For a case tested, the iterative replanning increased TCP from 0.87 for the repositioning plan to 0.995 for the adaptive voxel-based plan within 3 iterations, while all isodose line volumes in non-PTV regions were kept smaller or equal in size.
Conclusion: A method to quickly adapt a reference plan to a spatially changing tumor or response heterogeneity was proposed and demonstrated.
Contact Email: