Cloud-Based Monte Carlo Simulation Framework for Radiotherapy

Monte Carlo (MC) simulations are extensively used in radiotherapy to model the treatment head of clinical linear accelerators. However, the huge computation time, which can take weeks or even months, is the significant challenge researchers face. The high costs associated with setting up dedicated high-performance computing clusters further hinder accessibility. This study proposes a cloud-based Monte Carlo simulation framework leveraging the PRIMO program, deployed on Amazon Web Services (AWS), to address these challenges efficiently. PRIMO was installed on a Windows server-class machine hosted in the AWS cloud. The instance used was compute optimized (C5d.18xlarge) with 72 vCPUs, 144 GB memory, and 2 × 90 GB SSD storage. Making use of Amazon Elastic Compute Cloud (EC2) by Spot Requests method, considerably reduced the computational costs of MC simulation by utilizing a flexible and affordable pricing model. Cloud-based deployment demonstrated considerable improvements in scalability and efficiency. Simulations were made faster by the flexible allocation of cloud resources. In comparison with traditional on-premises hardware, cloud technology, reduced operational costs and enhanced accessibility for researchers. This study emphasizes the benefits of cloud computing for MC simulations in radiotherapy. By taking advantage of this method, PRIMO was efficiently deployed and executed. This method proved the feasibility of cloud-based MC simulations for high-performance applications. In this study, AWS is the only vendor considered. There is a further scope to include other vendors such as Microsoft Azure and Google Cloud, to compare costs and make it more cost-effective.