(Saturday, 4/7/2018) 10:30 AM - 12:30 PM
Room: Marquis Ballroom 5-8
Purpose: Significant uncertainties in commercial pencil beam algorithms for proton dose calculations demonstrate a need for an independent and accurate second check system. We developed a streamlined and automated workflow to perform Monte Carlo (MC) simulations of patient treatment plans.
Methods: A TOPAS based model of the HITACHI spot scanning nozzle, developed and commissioned earlier, was used to calculate dose and LET. Full simulation of the nozzle ensures suitable accuracy for collection of data for future retrospective studies on late effects and toxicities while also ensuring its utility as a second check of clinical plans. A framework was developed using MATLAB to automatically handle the processes such as data exchange between the TPS and the MC system, removal of CT artifacts, and inclusion of critical RT structures. DICOM RT plan information was used to generate TOPAS inputs describing the spot scanning pattern. Resultant dose files were processed and formatted as DICOM files for analysis. Patient simulations were performed using the institutional High Performance Computing Facility.
Results: With the new framework in place, we were able to save 30 – 60 minutes otherwise spent on manual processing of data. Most importantly, now any user can compare a patient-dose calculation from TPS and MC simulation without any knowledge of the complicated and previously manual intermediate steps.
Conclusion: Given the uncertainties in the pencil beam algorithms, it is valuable to use a MC method as an independent dose verification system. This work indicates the potential of the newly developed, automated MC second check system to be successfully implemented in the clinical workflow. This system will also be useful in generating MC data for retrospective studies. Although some of the methods we implemented to automate the patient simulations are specific to our institution, the process can be implemented in other proton therapy centers.
Contact Email
