Program Information
Water Equivalent Path Length Calculations Using Scatter-Corrected Head and Neck CBCT Images to Evaluate Patients for Adaptive Proton Therapy
J Kim*, Y Park , G Sharp , B Winey , Massachusetts General Hospital and Harvard Medical School, Boston, MA
Presentations
MO-FG-CAMPUS-JeP1-5 (Monday, August 1, 2016) 4:30 PM - 5:00 PM Room: ePoster Theater
Purpose: To establish a method to evaluate the dosimetric impact of anatomic changes in head and neck patients during proton therapy by using scatter-corrected cone-beam CT (CBCT) images.
Methods: The water equivalent path length (WEPL) was calculated to the distal edge of PTV contours by using tomographic images available for six head and neck patients received photon therapy. The proton range variation was measured by calculating the difference between the distal WEPLs calculated with the planning CT and weekly treatment CBCT images. By performing an automatic rigid registration, six degrees-of-freedom (DOF) correction was made to the CBCT images to account for the patient setup uncertainty. For accurate WEPL calculations, an existing CBCT scatter correction algorithm, whose performance was already proven for phantom images, was calibrated for head and neck patient images. Specifically, two different image similarity measures, mutual information (MI) and mean square error (MSE), were tested for the deformable image registration (DIR) in the CBCT scatter correction algorithm.
Results: The impact of weight loss was reflected in the distal WEPL differences with the aid of the automatic rigid registration reducing the influence of patient setup uncertainty on the WEPL calculation results. The WEPL difference averaged over distal area was 2.9 ± 2.9 (mm) across all fractions of six patients and its maximum, mostly found at the last available fraction, was 6.2 ± 3.4 (mm). The MSE-based DIR successfully registered each treatment CBCT image to the planning CT image. On the other hand, the MI-based DIR deformed the skin voxels in the planning CT image to the immobilization mask in the treatment CBCT image, most of which was cropped out of the planning CT image.
Conclusion: The dosimetric impact of anatomic changes was evaluated by calculating the distal WEPL difference with the existing scatter-correction algorithm appropriately calibrated.
Funding Support, Disclosures, and Conflict of Interest: Jihun Kim, Yang-Kyun Park, Gregory Sharp, and Brian Winey have received grant support from the NCI Federal Share of program income earned by Massachusetts General Hospital on C06 CA059267, Proton Therapy Research and Treatment Center.
Contact Email: