Encrypted login | home

Program Information

Seamless Integration of a GPU-Based Monte Carlo Treatment Plan Optimization Engine for Carbon Ion Therapy with Varian Eclipse System

no image available
M Tsai

M Tsai1*, X Jia2 , N Qin3 , (1) National Taiwan University, Taipei, Taiwan, (2) The University of Texas Southwestern Medical Center, Dallas, TX, (3) UT Southwestern Medical Center, Dallas, AA


SU-I-GPD-T-117 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: Monte Carlo (MC) simulation is widely regarded as the most accurate method for calculation of physics dose and biological dose in carbon ion therapy. It is however rarely used in the clinical inverse treatment planning process due to the extremely large computational burden and the required expertise. We have previously developed a novel system of GPU-accelerated MC-based treatment plan optimization. This abstract reports a novel integration approach to embed our system in Varian Eclipse treatment planning system to facilitate clinical translation of our system. Our study enabled fully MC-based biological dose optimization with a user-friendly operating environment and graphic user interfaces (GUI).

Methods: The system consists of a client-end served by the Eclipse planning environment and the server-end that is the actual MC-based optimization engine running on a remote GPU server. At the client end, a user can setup treatment planning parameters. After that, C# script program with Varian Eclipse Scripting Application Programming Interface (ESAPI) was used to acquire patient data and preprocess it, and transfer the planning request to the server-end. After that, our GPU-accelerated carbon therapy MC engine, goCMC, along with an integrated biological dose optimization module, was launched to perform MC simulation and inverse planning.

Results: We tested the system with three patient cases. Plan configurations including prescription dose, constraints for organs at risk (OAR) and field arrangement were set at Eclipse and sent to server-end along with patient imaging data. MC-based biological dose optimization was successfully performed with GPU acceleration for all testing cases.

Conclusion: By embedding goCMC and an biological optimization module into Eclipse, we have achieved a fast, user-friendly and fully MC-based treatment planning system for carbon ion therapy. This system will facilitate carbon ion therapy researchers and clinic.

Contact Email: