Temporal analysis of inter-fraction in-vivo 3D dosimetry using DVH difference curves based on daily exit fluence measures and cone-beam CT images
M. M. Zaini,1,2 G.A. Sandison2
1Northwest Medical Physics Center, Lynnwood, WA; 2University of Washington, Seattle, WA
ASTRO Annual Scientific Meeting, Boston, MA (2012).
Purpose/Objectives: Cancer patients undergoing highly fractionated radiation therapy may experience changes in the relative position of their tumor and surrounding normal tissues during the course of their therapy. These temporal changes impact the planned distribution of radiation, and it is vital that the attending radiation oncologist continually assess information on these changes to decide whether a modification to the patient’s treatment plan is warranted. The purpose of our study is to assist the attending in this assessment by providing 4D (temporal and spatial) patient dose distribution changes based on exit fluence measures overlaying CBCT images of the patient’s anatomy presented at the time of each treatment fraction.
Materials/Methods: Exit fluence was collected with the EPID during each treatment fraction and back-projected onto the daily acquired CBCT image volumes. Actual daily delivered dose was computed based on the daily CBCT by the Math Resolutions, DosimetryCheck® software. Any patient shift corrections based on the registration of the daily and planning CT volumes prior to administering the radiation treatments were incorporated in the 3D dose calculations. Processing of daily CBCT images involved the demarcation of the tumor. Daily DVH of the 3D in-vivo tumor dose was compared against that of the originally planned IMRT or VMAT treatment.
Results: Daily tumor DVH deviated from the planned DVH on different days. Subtraction of the daily DVH curves and the planned one was performed to generate a difference curve. The area under the planned DVH curve was then compared to this difference curve. The ratio of the absolute difference integrals and the planned DVH integral varied from 0.5% to 4.5%. The running sums of the absolute difference between the daily DVH curves and the planned DVH were also computed. Fluctuations in the daily DVH curves compared to the planned one did not result in target dose coverage changes by more than 5% in four of the patients studied. The running sum difference curves diminished in value over the course of treatment indicating that daily fluctuations in tumor DVH are more random than systematic in terms of dose coverage.
Conclusions: An attending may easily assess daily whether corrections to planned treatment are required based on DVH difference curves between planned and actual treatment dose distributions. Although the visual inspection of the slice-by-slice daily tumor coverage showed some positional dependence of the dose coverage deficiencies, it illustrated that the daily variation of dose is predominantly disseminated throughout the tumor. Absolute differences in daily DVH curves provides an accurate measure of daily variation in dose deposition but running sum differences may overestimate the biological impact of this anatomic-based treatment inaccuracy.