Study on the feasibility of quality improvement for automatic plans based on rapid plan model

Area of Science:

• Radiation Oncology
• Medical Physics
• Computational Biology

Background:

• Intensity-modulated radiation therapy (IMRT) planning relies on accurate models for optimal dose delivery.
• Developing interactive feedback systems can potentially improve the efficiency and quality of automated treatment planning.

Purpose of the Study:

• To evaluate the feasibility and effectiveness of an interactive internal feedback system for automatic radiation therapy planning.
• To compare the performance of two automatic planning models in generating IMRT plans for pelvic cancer patients.

Main Methods:

• Seventy pelvic patient cases were analyzed.
• Two automatic planning models (M1 and M2) were developed and trained iteratively using Intensity-modulated radiation therapy (IMRT) plans.
• Dosimetric parameters and model performance metrics were compared between the plans generated by M1 and M2.

Main Results:

• Model M2 demonstrated superior stability compared to M1.
• Key dosimetric parameters, including organ-at-risk sparing (e.g., bladder, femoral heads) and plan quality indices (e.g., D2, D98, HI), showed improvements with M2.
• The M2 model resulted in a significant reduction in monitor units (MUs).

Conclusions:

• The iterative
• plan-model-plan-model
• approach establishes a feasible and meaningful interactive internal feedback system.
• This system enhances the stability and quality of automatic IMRT plan generation.
• The findings support the integration of such feedback systems into clinical radiation oncology workflows.