Solution to the integrated optimization of dose, dose rate, and LET for proton FLASH therapy using a distributed parallel computing framework

Purpose: The recently proposed Integrated Physical Optimization IMPT (IPO-IMPT) framework has demonstrated the feasibility of simultaneously accounting for dose, dose rate, and linear energy transfer (LET) in patient treatment planning with proton FLASH. Here we present a solution to IPO-IMPT using a distributed parallel computing framework that was used to simulate the underlying quantum physics processes involved in radiation transport. We also present animal study simulations to demonstrate the need for doing Integrated Biological Optimization IMPT (IBO-IMPT). Methods: We have developed software which simultaneously optimizes the geometry of a patient-specific set of range compensating bars and range modulating pins, and the weights of a FLASH proton pencil beam spot map, for dose, instantaneous dose rate (IDR), and LET. The method uses the Open Science Grid (OSG), a distributed computing cluster, to efficiently complete the computationally intensive calculations. We also introduce the concept of a "restricted influence grid" as a technique for significantly improving computational performance. Additionally, animal study simulations were performed using TOPAS. Results: Our optimization technique created a sizable improvement in IDR and LET in the organs at risk (OARs) of our test patient with a negligible sacrifice to dose coverage of the CTV when compared to traditional IMPT. Our animal studies demonstrate that extra biological dose (XBD) should potentially be used to optimize for biological parameters in addition to physical parameters. Conclusion: This solution to IPO-IMPT is a promising tool for patient treatment planning which has the potential to be used to more effectively spare heathly tissue via the FLASH effect without sacrificing tumor killing efficacy.