Knowledge-based IMRT planning for individual liver cancer patients using a novel specific model

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• Intensity-modulated radiotherapy (IMRT) is a complex treatment for liver cancer.
• Special anatomical characteristics in patients pose challenges for IMRT planning.
• Knowledge-based planning systems aim to improve treatment efficiency and quality.

Purpose of the Study:

• To benchmark RapidPlan's performance against manual clinical plans for liver IMRT.
• To evaluate a specific RapidPlan model (Model-S) versus a general model (Model-G) for patients with unique anatomy.
• To assess the predictive capability of knowledge-based planning models.

Main Methods:

• Two RapidPlan models were developed: Model-S (30 patients with specific anatomy) and Model-G (60 patients).
• IMRT plans were generated using Model-G (RPG-plans), Model-S (RPS-plans), and manual planning (M-plans) for 13 evaluation patients.
• Dose-volume histograms (DVHs) were compared between plan types to assess plan quality and organ sparing.

Main Results:

• RapidPlan successfully generated clinically acceptable IMRT plans for all patients.
• RP-plans showed improved planning target volume (PTV) V95% and better right kidney sparing than M-plans.
• The specific model (RPS-plans) improved conformity index (CI) and spared the spinal cord and right kidney more effectively than the general model (RPG-plans), despite slightly higher normal liver dose.

Conclusions:

• RapidPlan is capable of producing high-quality IMRT plans for liver cancer.
• A specific knowledge-based model (Model-S) enhances PTV conformity and organ-at-risk (OAR) sparing in IMRT for individual patients.
• This approach offers improved treatment planning for liver cancer patients with challenging anatomies.