A retrospective study of automatic progressive optimization for lung cancer radiotherapy plans on the Halcyon and RayStation systems.

This study aims to evaluate the feasibility of using RayStation's scripting function to generate automated radiotherapy plans for non-small cell lung cancer (NSCLC) patients on a Varian Halcyon accelerator and to compare their dosimetric characteristics with those of retrospectively collected manual clinical plans. A total of 63 conventional fractionation plans for NSCLC, previously designed using RayStation 4.5 for a variety of linear accelerators-including Trilogy, TrueBeam, Halcyon, and Elekta Infinity-were compared with automated plans generated using RayStation 9.0 for Halcyon. This heterogeneous control group was chosen to reflect real-world clinical practice across multiple platforms. Target coverage, doses to organs at risk (OARs), monitor units, and plan complexity were assessed. The automated plans showed improved dose conformity and lower OAR exposure under the planning configuration used. However, these differences should be interpreted with caution, as the comparison involved different treatment planning systems (TPS) versions and hardware platforms. Further controlled studies using the same TPS and linac are needed to validate the observed improvements.