Premium
SU‐E‐J‐106: The Use of Deformable Image Registration with Cone‐Beam CT for a Better Evaluation of Cumulative Dose to Organs
Author(s) -
Fillion O,
Gingras L,
Archambault L
Publication year - 2015
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4924193
Subject(s) - contouring , nuclear medicine , voxel , cumulative dose , medicine , dosimetry , radiation treatment planning , image registration , radiation therapy , randomized controlled trial , cone beam computed tomography , medical imaging , radiology , computed tomography , computer science , surgery , image (mathematics) , artificial intelligence , computer graphics (images)
Purpose: The knowledge of dose accumulation in the patient tissues in radiotherapy helps in determining the treatment outcomes. This project aims at providing a workflow to map cumulative doses that takes into account interfraction organ motion without the need for manual re‐contouring. Methods: Five prostate cancer patients were studied. Each patient had a planning CT (pCT) and 5 to 13 CBCT scans. On each series, a physician contoured the prostate, rectum, bladder, seminal vesicles and the intestine. First, a deformable image registration (DIR) of the pCTs onto the daily CBCTs yielded registered CTs (rCT) . This rCT combined the accurate CT numbers of the pCT with the daily anatomy of the CBCT. Second, the original plans (220 cGy per fraction for 25 fractions) were copied on the rCT for dose re‐calculation. Third, the DIR software Elastix was used to find the inverse transform from the rCT to the pCT. This transformation was then applied to the rCT dose grid to map the dose voxels back to their pCT location. Finally, the sum of these deformed dose grids for each patient was applied on the pCT to calculate the actual dose delivered to organs. Results: The discrepancy between the planned D98 and D2 and these indices re‐calculated on the rCT, are, on average, of −1 ± 1 cGy and 1 ± 2 cGy per fraction, respectively. For fractions with large anatomical motion, the D98 discrepancy on the re‐calculated dose grid mapped onto the pCT can raise to −17 ± 4 cGy. The obtained cumulative dose distributions illustrate the same behavior. Conclusion: This approach allowed the evaluation of cumulative doses to organs with the help of uncontoured daily CBCT scans. With this workflow, the easy evaluation of doses delivered for EBRT treatments could ultimately lead to a better follow‐up of prostate cancer patients.