A knowledge-based approach to automated planning for hepatocellular carcinoma

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• Intensity modulated radiation therapy (IMRT) planning is complex and time-consuming.
• Automated planning systems aim to improve efficiency and consistency in radiation therapy.
• Knowledge-based models leverage past treatment data to optimize future plans.

Purpose of the Study:

• To develop a knowledge-based model for liver cancer Auto-Planning within the Pinnacle treatment planning system.
• To evaluate the effectiveness of automated inverse intensity modulated radiation therapy (IMRT) planning compared to manual methods for liver cancer.

Main Methods:

• Dose-volume histogram (DVH) data from 50 Tomotherapy patients were used to build the Auto-Planning model.
• A separate cohort of 20 patients underwent both manual and automated planning for comparative analysis.
• Treatment plans were evaluated blindly for target coverage, organ at risk (OAR) doses, and planning time.

Main Results:

• Automated plans demonstrated superior target conformity index (CI) and significantly reduced doses to normal liver, left kidney, and small bowel.
• Mean target dose was slightly higher (0.5 Gy) in automated plans, but target homogeneity index (HI) differences were not significant.
• Automated planning substantially decreased overall working time compared to manual planning.

Conclusions:

• The developed knowledge-based Auto-Planning model is effective for liver cancer treatment.
• Automated planning offers significant advantages in efficiency and organ sparing for liver cancer patients.
• This approach shows promise for optimizing automated inverse intensity modulated radiation therapy (IMRT) delivery.