Development and clinical implementation of a comprehensive multifractionation scheme HyperArc-based RapidPlan model for single and multiple brain metastases

Propose a solution to help mitigate the 1 to 2-week delay from simulation to treatment time for single-isocenter single- and multi-lesion brain stereotactic radiosurgery/radiotherapy (SRS/SRT) plans. Utilizing new machine learning techniques to automate and standardize the treatment planning process. 89 previously treated high-quality, highly conformal HyperArc SRS/SRT plans delivered on a TrueBeam LINAC (6MV-FFF), were used to develop a HyperArc-based RapidPlan (HARP) model. The training set consisted of 69 plans (8 single-fraction, 15 to 22 Gy; 33 three-fraction, 24 to 27 Gy; 28 five-fraction, 25 to 40 Gy) with a total of 133 metastases. The testing set consisted of 20 plans (3 single-fraction, 9 three-fraction, 8 five-fraction) with a total of 45 metastases. The model was trained then the testing plans were replanned using the HARP model. The HARP model was then used clinically to treat our first patient. Across all 3 fractionation schemes, the HARP model produced clinically acceptable plans in under 20 minutes. For target coverage, the HARP plans achieved a comparable PTV D_95% with a median of 101.48% compared to the manual plan's 101.33%. The GTV D_100% received a median of 108.24% and 105.39% for the HARP plan and original plan, respectively (p < 0.01). Maximum doses to organs-at-risk (OARs) were well within standard SRS/SRT criteria. All plans received clinically acceptable EPID-based portal dosimetry gamma pass rates and Monte Carlo 2nd check rates. The clinical 2-lesion HARP plan was generated in 15 minutes and used for patient treatment. The comprehensive HARP model can handle 3 different fractionation schemes, adequately sparing OARs, including normal brain toxicity, delivering a therapeutic dose to the brain lesions, generating plans in under 20 minutes. The utilization of HARP models can significantly decrease the treatment planning time, allowing institutions to standardize patient care, optimize the clinical workflow, and increase the accessibility of high-quality radiation therapy treatments.