University of Maryland School of Medicine, Baltimore Ellicott City, MD
W. Yao1, B. Zhang1, M. Ferris1, J. K. Molitoris1, M. Huang2, N. C. Biswal1, and B. Y. Yi1; 1Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD, 2Mayo Clinic, Jacksonville, FL
Purpose/Objective(s): Adaptive plans are often required for proton therapy due to the patient’s anatomy change. Scheduled weekly or bi-weekly CT rescans (QACTs) to filter the significant change have been a standard practice, but much increase the workload of staff and patients CT dose. In our previous study, we obtained a plan-independent range tolerance such that no replan was done when the range change in CBCT was less than the tolerance. 56% of QACTs were determined skippable. This work is to further improve prediction accuracy by developing a plan-specific range tolerance (PSRT). The aim of this study is to evaluate the PSRT in reducing the number of QACTs and predicting adaptation. Our hypothesis is that PSRT with daily CBCT can replace scheduled QACT to catch the timepoint of adaptation. Materials/
Methods: We studied the quantitative relation between dose and range in the treatment plan. This novel study led to obtaining PSRTs corresponding to the target D95=95% and V98=90%, two of the adaptation criteria, in the planning CT. The PSRTs were then applied to the daily CBCTs of the patient treated with the plan. We tested PSRT retrospectively in 58 non-nasal head and neck patients and the max dose <110% of the prescription in the worst-case scenario from April to December 2023, followed by the pilot study including all 20 patients with the same exclusions in January and February 2024. There were 183/85 QACTs, 16/6 adaptations and 1411/498 CBCTs associated with the retrospective/pilot study cases. For each CBCT, the water equivalent thickness (WET) along the pencil beam path between the target surface and the patient body was calculated. The WET of the first day CBCT was used as the reference, and the WET change (dWET) in each CBCT from the first day CBCT was used as the surrogate for proton range change. If on a QACT day, dWET > PSRT in the CBCT and an adaptation is determined necessary from the QACT, our test is true, and otherwise false. Results: For the retrospective cases, 83% (152 out of 183) of the scheduled QACTs were skippable. For the pilot study cases, this number was 80% (68 out of 85). If we triggered a QACT once dWET > PSRT in the pilot study, 73 QACTs would be triggered. If we triggered a QACT only when dWETs > PSRT for 2 consecutive days, 38 QACTs would be required. Namely, 45% (38 out of 85) of QACTs were required, and meanwhile the timepoint for adaptation was better caught than the scheduled QACT method because the CBCT was performed daily. Compared to the replan decisions from the scheduled QACT, 3 cases showed false negatives in the retrospective cases and 2 cases in the pilot study: 4 of them were due to bad setup during QACT (no IGRT in the CT room) and the other due to a hot spot (110%) in the air cavity. However, for these 4 missed adaptations due to bad setup, dWET > PSRT within 2 days. Namely, the adaptation would not be missed for more than 2 days. The specificity and sensitivity were 86% and 67%. Conclusion: Plan-specific range tolerance with CBCT can be used to predict the adaptive plan and significantly reduce the unnecessary number of QACTs.
