Develop a novel multi-approach dynamic Spot Scanning Proton Arc optimization framework

About This Grant

Project Summary Spot-scanning Proton Arc (SPArc) therapy is an emerging treatment modality that has shown potential clinical benefits for a wide range of disease sites. Unlike the conventional step-and-shoot multi-field Intensity Modulated Proton Therapy (IMPT), SPArc therapy utilizes arc trajectories for planning and treatment. The extra degree of freedom offers superior dose conformity which is particularly crucial for treating head and neck squamous cell carcinoma (HNSCC) patients, where organs-at-risk (OARs) are often adjacent to the target volume. Therefore, sharp dose fall-off using SPArc therapy is able to reduce radiation-induced toxicity. In 2018, the concept of SPArc therapy was successfully demonstrated in a clinical Proton Therapy System (PTS) equipped with a synchrocyclotron accelerator and partial gantry. This breakthrough drew significant interest from the radiation oncology community. However, the clinical implementation of SPArc therapy faces several challenges. (1) Slow optimization speed: Because of the mathematical complexity and optimization burden associated with numerous spot and energy layers, it takes hours to optimize a treatment plan; (2) Prolonged treatment delivery time: Various PTS have different machine-specific delivery sequences (MSDS) such as synchrotron, cyclotron, and synchrocyclotron accelerators. It makes it almost impossible to generate a universal delivery-efficient plan across all the different PTS. Some treatment times could be doubled because the key plan parameter does not fit the specific type of PTS even with an identical plan that worked for another PTS; (3) Dose errors in the dynamic treatment delivery: The accuracy of proton beam therapy is heavily reliant upon beam angle and its associated water-equivalent-path because of the sharp distal fall-off. Delivering the dose continuously during gantry rotation could cause unnecessary dose error between the "true" dynamic delivery and the static nominal plan, leading to the degradation in local tumor control probability by about 5% for the HNSCC patient population. To overcome these impediments and enable the clinical implementation of SPArc therapy on a global scale, this project will develop a novel multi-approach dynamic SPArc optimization framework that ensures fast optimization, efficient treatment delivery, and accurate dosimetry for various types of PTS. The project includes three specific aims: (Aim 1) Systematically investigate machine-specific delivery sequences of various PTS and establish SPArc simulators; (Aim 2) Develop a fast sparsity optimization framework; and (Aim 3) Introduce “true” dynamic sequencing SPArc optimization framework and assess the improvement in the treatment delivery accuracy. The results of this study will provide the first-of-its-kind SPArc delivery simulator and an open-source SPArc sparsity optimization framework, benefiting global investigators and researchers and facilitating prospective clinical trials.