Validation of knowledge-based and multicriterial optimization assistive auto-planning algorithms for prostate VMAT...

External beam radiotherapy for cancer treatment historically has faced the challenge of delivering sufficient dose to the target while minimizing dose to critical organs. Inverse planning techniques in modulated therapy can improve organ-at-risk (OAR) sparing but require significant human resources and can depend on planner experience. Advances in software and artificial intelligence (AI) have enabled the development of commercial treatment planning systems (TPS) with scripting and auto-planning capabilities, potentially reducing human resource demands and standardizing plan quality. This study aimed to create and validate knowledge-based planning (KBP) and multicriterial optimization (MCO) algorithms for assistive auto-planning of prostate volumetric modulated arc therapy (VMAT) plans using the Elekta Monaco TPS.

Our methodology involved implementing and validating these algorithms retrospectively. We compared dose volume histogram (DVH), generalize equivalent uniform dose (gEUD), and plan quality scores with KBP and MCO algorithms to the clinical plans.

KBP and MCO generated plans on average spared OARs better and on average had better conformity to the targets whilst not sacrificing target coverage. Additionally, algorithm generated plans could be produced within 30 min and were of clinical quality for 72% (KBP) and 78% (MCO) of plans. When comparing the KBP and MCO plans, MCO was dosimetrically slightly superior, slightly faster, and produced plans of clinical quality in more of the validation population than the KBP algorithm.

This work showed that it is possible to build KBP and MCO based assistive auto-planning in the Elekta Monaco TPS focusing on gEUD based cost functions. When using the Monaco Scripting functionality, these algorithms could assist in reducing the clinical planning workload while maintaining a patient specific standard of quality.