Designing Robust and Automated Treatment Plans Utilizing Dynamic Couch Motion With Automated Treatment Planning: Optimizing Couch Motion for Precise Delivery ― Part A

Abstract

<h3>Purpose/Objective(s)</h3> This study finds a relationship among couch speed, dose rate, and MLC, allowing gradual start and stop positioning of the couch for treatments utilizing dynamic couch-based dose delivery at static gantry positions. <h3>Materials/Methods</h3> We introduce a gantry static couch motion (GsCM) based technique, which delivers the dose using a dynamic couch and static gantry positions. It is imperative to study couch motion behavior for techniques like GsCM, where the patient is a moving target throughout the dose delivery. Sudden changes in couch speed can result in a shift in position and discomfort for the patient. Therefore, it is necessary to implement a smooth start, a constant speed in the middle, and deceleration at the couch's stopping positioning. We studied the relationship among the following parameters: dose rate, MLC speed, and couch motion. To fulfill this, we simulated a single treatment arc manually in the treatment planning system that includes multiple couch positions at static gantry positions. All the fields were merged into a single deliverable arc while exporting into the Extensible Markup Language (XML) format using the TPS scripting API. Developer mode on a Linac performed the treatment delivery. The second part of this study looked into the various ways to control the couch motion. Methods tried included a fixed couch speed versus controlling the couch speed by limiting the dose rate. Trajectory log files helped analyze the couch positional error (RMSE). <h3>Results</h3> We found couch speed optimization by limiting the dose rate is better than the MLC speed or couch speed. A slow ramping up and down of the speed with a constant speed in the middle of an arc field is achievable. Linac log file analysis showed minimum couch position error (RMSE: median 0.035⁰, range 0.02-0.0.045⁰) for 1° couch speed as compared to system defined couch speed (RMSE: median 0.07⁰, range 0.05-0.12⁰), and (RMSE: median 0.07⁰, range 0.06-0.09⁰) for 2° couch speed respectively. Even though 1° couch speed provided a minimum couch RMSE error but at the expense of a longer treatment time which is not suitable for keeping the patient longer on the treatment table. For the dose rate limiting option, couch positioning error at the start and stop location (RMSE: median 0.008⁰, range 0.006-0.02⁰) showed a gradual and accurate ramping up and down of the couch, and at the middle of the arc remained the same as noticed for the system defined couch speed (RMSE: median 0.09⁰, range 0.065-0.12⁰). <h3>Conclusion</h3> This feasibility study serves as an end-to-end test and is probably the first of its kind that thoroughly validates various ways to control the couch speed and its integration with patient motion during the dose delivery. Treatments utilizing dynamic couch-based dose delivery must include patient immobilization. Our study takes patient motion into account and provides gradual start and stop motions of the couch, which helps minimize the patient motion for treatments utilizing dynamic couch motion.