Program Information
Assessing the Need and Feasibility for Online Plan Adaptation Based On Daily CBCT of Head and Neck Proton Therapy Treatments
P Botas1,2*, J Kim2 , C Grassberger2 , B Winey2 , H Paganetti2, (1) Heidelberg University, Heidelberg, Germany (2) Massachusetts General Hospital and Harvard Medical School, Boston, MA
Presentations
TH-AB-605-2 (Thursday, August 3, 2017) 7:30 AM - 9:30 AM Room: 605
Purpose: To demonstrate the necessity and feasibility of online Monte Carlo (MC) dose-of-the-day calculation on Cone Beam CTs (CBCTs) studying a head and neck patient cohort. Dose-of-the-day evaluation is the first step towards adaptive IMPT.
Methods: IMPT plans of 10 head and neck cancer patients were evaluated on 6-7 weekly CBCTs per patient. A previously developed a-priori CT-based scatter correction technique showing ~1.9% range uncertainty was employed to process the CBCTs. The plan was then simulated in the weekly patient geometry using a GPU MC code, gPMC. The contours were propagated from the CT to the CBCTs using B-spline Deformable Image Registration (DIR). Dosimetric indices of target and Organs at Risk (OARs) were studied in the subsequent fractions and compared to the initial plan. All doses are reported assuming the whole treatment was delivered to the fraction geometry.
Results: The total evaluation time per fraction was <10min for all patients, 6min for processing the CBCT and 1 min for dose calculation and DIR. The mean dose to CTV decreased from an average across patients of 66.5±1.8Gy to 65.2±1.9Gy at the 6th fraction. In the same period, the average V100 decreased from 99.0% to 93.1±5.7%. In 2.5 out of 6 fractions per patient, V100 was below 95%. The dose heterogeneity (D5-D95) increased from 5.9 to 12.3Gy. The mean dose increased on average from 8.8 to 12.7Gy in the oral cavity, and from 8.3 to 11.2Gy in the larynx. The maximum mean dose changes in a specific patient were +5.1/+5.8Gy in the oral cavity/larynx.
Conclusion: A framework to assess the feasibility of online adaptation (plan-of-the-day) within clinical time constraints using a GPU MC dose calculation algorithm was developed. Significant plan degradation was demonstrated, reducing tumor coverage and OAR sparing. This work enables re-evaluation of plan constraints and objectives, potentially triggering plan adaptation.
Funding Support, Disclosures, and Conflict of Interest: Research supported by NIH/NCI U19 CA-21239
Contact Email: