Quality and mechanical efficiency of automated knowledge-based planning for volumetric-modulated arc therapy in head and neck cancer

Abstract

Objectives: This study aimed to examine the effectiveness of the automated RapidPlan in assessing plan quality and to explore how beam complexity affects the mechanical performance of volumetric modulated arc therapy for head and neck cancers. Materials and methods: The plans were first generated using automated RapidPlan with scripting application programming interface (API) and then further refined through manual optimization (RP+MP) to improve plan quality. The quality of 20 plans was assessed, taking into account dose statistics and clinical plan acceptability. The impact of beam complexity on mechanical performance was analyzed using parameters such as leaf speed (LS), leaf acceleration (LA), mean-field area (MFA), cross-axis score (CAS), closed leaf score (CLS), small aperture score (SAS), and monitor units per control point (MU/CP). Patient-specific quality assurance (PSQA) was conducted to determine differences between the RP+MP and original plans. Results: No differences in the heterogeneity index and conformity number were observed between the RP+MP and original plans. The RP+MP plan was superior to the original plan for sparing the left cochlea, left and right internal auditory canals, chiasm, and left optic nerve. Significant differences (p < 0.05) were identified in CAS, SAS1 mm, SAS2 mm, and SAS10mm. However, there was no significant difference in PSQA between the RP+MP and original plans. The RP+MP plan without any modifications was clinically acceptable in 45% of cases. Conclusion: The automated RP with scripting API followed by MP (RP+MP) yielded a high-quality plan in terms of dose statistics and clinical acceptability. The RP+MP plan yielded a higher CAS and SAS compared with the original plan. Nevertheless, there was no significant difference in PSQA between the RP+MP and original plans.