Encrypted login | home

Program Information

Development of a Fast Monte Carlo Dose Calculation System for Online Adaptive Radiation Therapy Quality Assurance

Y Wang

Y Wang*, T Mazur , J Park , D Yang , S Mutic , H Li , Washington University School of Medicine, St. Louis, MO


TU-D-205-3 (Tuesday, August 1, 2017) 11:00 AM - 12:15 PM Room: 205

Purpose: Online adaptive radiation therapy based on real-time magnetic resonance imaging represents a paradigm-changing treatment scheme. However, conventional patient-specific quality assurance (QA) method based on phantom measurements is not feasible with the patient on the treatment couch. The purpose of this work is to develop a fast Monte Carlo (MC) system, gDPMvr, for validating online re-optimized tri-60Co IMRT adaptive plans with both high accuracy and speed.

Methods: The Monte Carlo system is based on DPM with further simplification of electron transport and consideration of external magnetic fields. The vendor-provided head model was incorporated into the code. Both GPU acceleration and variance reduction were implemented. Our previously developed dose calculation system gPENELOPE has been validated to be as accurate as the original PENELOPE code with significantly improved efficiency. We thus consider gPENELOPE as the standard. Four different types of phantoms were used to compare the accuracy and performance of the two DPM-based codes, gDPM and gDPMvr, to gPENELOPE, including a homogenous water phantom, a water-lung-water phantom, a water-lung-tumor-water phantom, and 15 clinical IMRT plans including stomach (4), lung (2), liver (3), adrenal gland (2), pancreas (2), spleen (1), and mediastinum (1).

Results: Percentage depth dose and off-axis dose profile comparisons in the synthetic phantoms demonstrate excellent agreements between gPENELOPE, gDPM and gDPMvr. For 15 clinical IMRT plans, gDPMvr reached 1% local dose uncertainty within 2.3 minutes on average.

Conclusion: With the imminent introduction of MRI-guided linac devices, integrating variance reduction techniques on GPUs will be a challenging and pressing problem. Moreover, the required accuracy of an online MC system and the appropriate QA metrics, e.g. conventional gamma criteria vs. dose-volume-histogram values based metrics, must be established. Nonetheless, gDPMvr achieves speeds beyond those required by in-use workflows for MRgRT ART while preserving accuracy comparable to that achieved by more traditional MC code.

Contact Email: