Room: Karl Dean Ballroom C
Purpose: To investigate the feasibility of adaptive dose painting by number (DPbN) planning using a novel object of dose prescription function (DPF) derived from multiple 18F-FDG PET/CT images.
Methods: Adaptive DPbN plans with hybrid fractionation schemes were designed and evaluated retrospectively for 6 head-and-neck cancer (HNC) patients (3 of 6 were locoregional failures) with human papillomavirus (HPV) status negative. For the first 20 of 35 fractions, a conventional plan with a uniform prescription dose (1.54Gy/fraction to CTV; 2Gy/fraction to GTV) was applied. For the remaining 15 fractions, an adaptive DPbN plan with the same prescription dose to CTV but a non-uniform prescription dose to GTV (1.54Gy~5.3Gy/fraction to each tumor voxel) was applied. The non-uniform prescription dose was obtained from a pre-constructed DPF with the desired TCP of 90% and low/high constrains of 54/150Gy biological equivalent dose in 2Gy (EQDâ‚‚). The adaptive DPbN plans were optimized using the DPF as the object and same normal organ constraints and compared to the conventional plans with the uniform prescription dose in 35 fractions.
Results: For the adaptive DPbN plans, tumor sub-volumes with prescription vs planned dose â‰¥ 3Gy/fraction were (2.78Â±3.05)cc vs (2.52Â±2.78)cc and the fraction of tumor voxels in GTV have a planned dose â‰¥ 95% of prescription dose was (95.86Â±3.41)%. Compared to the conventional plans, the DPbN plans reduced the mean dose of (1.3Â±18.83)% and (5.57Â±4.29)% for left and right parotid glands and the maximum dose of (3.45Â±9.77)%, (10.24Â±11.57)% and (4.24Â±9.16)% for spinal cord, brain stem and mandible respectively.
Conclusion: The adaptive DPbN plan with hybrid fractionation schemes can achieve the desired dose escalation with significant TCP improvements by tailoring high fractional doses to the resistant tumor voxels while keeping the radiation to critical structures below or similar to those of the conventional plan for HNC patients with HPV status negative.
Funding Support, Disclosures, and Conflict of Interest: Sponsored partially by the research software support from Philips Radiation Oncology Systems
Not Applicable / None Entered.