Knowledge-based intensity-modulated proton planning for gastroesophageal carcinoma

Area of Science:

• Radiation Oncology
• Medical Physics
• Oncology

Background:

• Intensity-modulated proton therapy (IMPT) offers precise dose delivery for complex cancer cases.
• Optimizing IMPT plans, especially for gastroesophageal junction tumors, requires sophisticated tools.
• Knowledge-based planning (KBP) systems aim to streamline and improve treatment plan quality.

Purpose of the Study:

• To evaluate the performance of a narrow-scope, knowledge-based RapidPlan (RP) model for IMPT in locally advanced gastroesophageal junction carcinoma.
• To compare RP-optimized IMPT plans against manually optimized plans.
• To assess the predictive accuracy and clinical applicability of the RP model.

Main Methods:

• A retrospective cohort of 60 patients with locally advanced gastroesophageal junction carcinoma was analyzed.
• 45 patients' data were used to train a dose-volume histogram (DVH) predictive model within the RP system.
• The remaining 15 patients' data served for independent validation, comparing RP plans to manual plans using quantitative dose-volume metrics.

Main Results:

• Both manual and RP-optimized IMPT plans achieved similar dosimetric outcomes, meeting all planning objectives.
• For the validation set, RP plans showed comparable or slightly improved target coverage (e.g., PTV V98%: 91.4% vs 89.3%) and homogeneity.
• No significant differences in dose to organs at risk (lungs, heart, LAD artery, kidneys, spleen, spinal canal) were observed between the two planning methods.

Conclusions:

• A narrow-scope knowledge-based RP model is effective for optimizing IMPT for gastroesophageal junction cancer.
• RP-based optimization demonstrated equivalence to manual planning, offering a reliable alternative.
• The study highlights the predictive power of the RP method, with high correlation between predicted and achieved doses.