Encrypted login | home

Program Information

A Fully Monte Carlo-Based Efficient Biological Treatment Plan Optimization System for Intensity Modulated Carbon Ion Therapy

N Qin

N Qin1*, C Shen1 , M Pinto2 , Z Tian1 , G Dedes2 , A Pompos1 , S Jiang1 , K Parodi2 , X Jia1 , (1) University of Texas Southwestern Medical Center, Dallas, TX, (2) Ludwig-Maximilians-Universitat Munchen, Garching, Bavaria


TH-AB-605-1 (Thursday, August 3, 2017) 7:30 AM - 9:30 AM Room: 605

Purpose: Biological effects must be considered in carbon ion therapy treatment planning. For intensity modulated carbon therapy (IMCT), it is desirable to employ Monte Carlo (MC) methods to compute properties of each pencil beam spot for treatment planning because of its accuracy in modeling physics process and estimating biological effects. We have previously developed goCMC, a GPU-oriented fast cross-platform MC engine for carbon ion therapy. The purpose of this study is to build a biological treatment plan optimization system on top of goCMC.

Methods: Repair-Misrepair-Fixation (RMF) model was integrated into goCMC to compute spatial distribution of linear-quadratic model parameters for each spot. Other radiobiological models are supported by using user-provided data tables. Based on prescribed biological dose distribution, we computed corresponding logarithm of cell survival log(S). Our treatment plan optimization module then minimizes difference between the prescribed and actual log(S). The choice of using log(S) simplified the optimization problem compared to the typical approach of minimizing difference between prescribed and actual biological dose. A gradient-based algorithm was employed to solve the optimization problem. We tested our system in 1-dimensional water phantom and 3-dimensional (3D) patient cases.

Results: In all cases, our system was able to generate treatment plans with biological spread-out Bragg peaks covering the targeted regions while sparing critical structures. Using one NVidia Geforce GTX Titan-Black GPU, total computation time including spot simulation and optimization was 1.7 hours for the tested prostate case (8282 spots), and 0.5 hours for the pancreas case (3795 spots).

Conclusion: We have developed a fully MC-based biological treatment plan optimization system for IMCT. MC simulation was performed by a GPU-oriented simulation engine, goCMC. With the built-in RMF model or user-provided radiobiological models, the system is able to calculate and optimize biological dose for 3D patient cases in a clinically viable time frame.

Contact Email: