Anatomical dose tracking for adaptive radiation therapy
M. M. Zaini,1,2 W. B. Jackson,1,2 K. K. Thuo,1 G. A. Sandison,3 H. P. Patel,1,2 M. T. Luckstead,1,2 M. A. Whiton,2 L. E. Sweeney1
1Northwest Medical Physics Center, Lynnwood, WA; 2Skagit Valley Regional Health Center, Mount Vernon, WA; 3University of Washington, Seattle, WA
ASTRO Cancer Imaging and Radiation Therapy Symposium, Atlanta, GA (2011).
Purpose/Objective(s): To introduce a method that allows for inter-fraction modifications to a treatment plan so as to compensate for tumor shrinkage, patient weight loss, and anatomical shifts during a radiation treatment course. This method may lead to reduced morbidity and improved survival for certain patients compared to a therapy course designed using treatment plans produced days or weeks before treatment start.
Materials/Methods: During the design of a radiation treatment plan, tissue volumes of interest (VOIs) are delineated on CT-simulation images. These delineated initial VOIs that are transferred from the planning-CT to cone-beam CT (CBCT) image volumes serve as the starting point for creating the corresponding VOI structure sets in the CBCT images. In addition, CBCT volumes are registered to the planning-CT volumes by the therapists on the daily basis. The daily change in shape, volume, and position of the VOIs is determined by employing elastic mutual information technique. An electronic portal imaging device (EPID) used for acquiring daily CBCT images of the patient also allows collection of exit fluence images from the patient during the treatment. Actual daily 3D dose distribution delivered to the patient may then be computed from these measured exit fluences by back-projection onto the CBCT volumes. Comparison of the 3D dose distribution in the daily CBCT volumes with each other and the planning-CT allows the attending radiation oncologist to decide whether any change to the course of treatment is required. Temporally integrated dose to each individual voxel is needed in order to assess the tumor control probability and normal tissue complication probability due to damage of organs at risk (OARs), especially serial type OARs. Optical flow and fluid dynamic models are employed to quantify the temporally evolving target and OAR dose voxels.
Results: Ten patients with various cancers were selected for this study. Visualization of the results included velocity vector maps, time-dose curves, and movie techniques. Noticeable anatomical and dosimetric changes were observed in some patients, especially for lung and head and neck treatments. On the basis of these observations, it is estimated that at least 10% and possibly 40% of patients will benefit from inter-fraction modification of radiation treatment plans.
Conclusions: Daily CBCT dosimetry using measured exit fluence provides a direct and non-invasive means of determining anatomical imparted dose to VOIs. This information is valuable to assess whether OAR dose needs to be reduced or cold spots in the target corrected. Modification of treatment plans is warranted based on this limited study; however, the technique used to evaluate the importance of adaptive radiation therapy requires more extensive study.