Related Experiment Videos
Knowledge-based treatment planning: An inter-technique and inter-system feasibility study for prostate cancer
Elisabetta Cagni1, Andrea Botti1, Renato Micera2
1Medical Physics Unit, Department of Advanced Technology, Arcispedale Santa Maria Nuova, IRCCS, Reggio Emilia, Italy.
Summary
Knowledge-based (KB) RapidPlan models, trained on Helical Tomotherapy (HT) plans, successfully generated clinically acceptable prostate cancer treatment plans using different techniques and treatment planning systems. This demonstrates the utility of KB models for new treatment modalities.
Area of Science:
- Radiation Oncology
- Medical Physics
Background:
- Knowledge-based planning (KBP) offers potential for improving treatment plan quality and consistency.
- Helical Tomotherapy (HT) is an established radiotherapy technique.
- RapidPlan is a KBP system used to create treatment planning models.
Purpose of the Study:
- To evaluate the feasibility and performance of RapidPlan knowledge-based (KB) models trained on HT plans.
- To generate prostate cancer treatment plans using different techniques (RapidArc) and treatment planning systems (Tomoplan) with KB models.
- To validate the clinical acceptability of KB-generated plans against expert-created plans.
Main Methods:
- Two sets of prostate cancer cases (low risk and intermediate risk) treated with HT were used to train RapidPlan KB models.
- KB models were used to generate RapidArc (KB-RA) plans for inter-technique validation and HT plans in Tomoplan (KB-HT) for inter-system validation.
- KB plans were compared to manually created expert planner (EP) plans.
Main Results:
- RapidPlan models were successfully configured using HT plans.
- KB-RA plans achieved 100% and 92% fulfillment of dose-volume requirements for planning target volumes (PTVs) and organs at risk (OARs), respectively.
- KB-HT plans showed slightly lower fulfillment rates (90% for PTVs, 86% for OARs). KB-RA plans resulted in higher bladder and rectum doses compared to EP plans, while KB-HT and EP plans showed similar outcomes.
Conclusions:
- RapidPlan can be trained using plans from one treatment modality to create models for others.
- These KB models are suitable for generating clinically acceptable plans across different techniques and treatment planning systems.
- KB models based on established techniques can be beneficial when implementing new treatment modalities.