Multiscale spatial relationship-based model for predicting bladder wall dose in pelvic radiotherapy.

Area of Science:

• Radiation oncology
• Medical physics
• Cancer treatment

Background:

• Radiotherapy for pelvic tumors can lead to bladder toxicity.
• Accurate bladder wall dosimetry is crucial for treatment planning.
• Minimizing radiation dose to the bladder is essential for patient outcomes.

Purpose of the Study:

• To develop a prediction model for bladder wall dosimetry in pelvic tumor radiotherapy.
• To facilitate refinement and evaluation of radiotherapy treatment plans.
• To mitigate bladder toxicity in patients undergoing pelvic radiation therapy.

Main Methods:

• Collected radiotherapy treatment plans from rectal and gynecologic cancer patients.
• Utilized multiple linear regression to create prediction models based on spatial relationships between planning target volume (PTV) and bladder.
• Validated model accuracy in a cohort of new patients and used predictions for treatment plan optimization.

Main Results:

• Models developed from individual disease data showed superior performance.
• Predicted bladder wall dose parameters demonstrated high accuracy, with a correlation coefficient of 0.943.
• Treatment plan optimization using predicted values significantly reduced bladder wall dose (mean reduction of 2.27 Gy for V5.8% and 2.96 cm³ for V7.9%).

Conclusions:

• The developed prediction model is a valuable tool for predicting and minimizing bladder wall dose.
• This approach aids in optimizing and evaluating radiotherapy treatment plans for pelvic tumors.
• The model shows promise for reducing bladder toxicity and improving patient outcomes in radiation oncology.