Optimal radiotherapy dose scheduling with variable fraction sizes and breaks via sequential mixed-integer convex...

Area of Science:

• Radiation Oncology
• Mathematical Modeling
• Cancer Biology

Background:

• Standard radiotherapy uses equi-dose fractions, but non-uniform schedules may enhance tumor control.
• Tumor response is influenced by cellular resource competition and hypoxia.

Purpose of the Study:

• To develop an optimization method for constructing non-uniform radiotherapy dose schedules.
• To improve tumor control probability (TCP) by optimizing fraction size and treatment breaks.

Main Methods:

• Formulated dose scheduling as a finite-horizon optimal control problem.
• Employed a convex-concave procedure to solve the nonconvex problem with constraints on normal tissue dose and daily limits.
• Calibrated a tumor dose-response model to head-and-neck squamous cell carcinoma.

Main Results:

• Optimal schedules featured an initial "primer shot", a 1-week break, followed by smaller fractions and a final large fraction.
• The schedule leverages reoxygenation of hypoxic cells during the break for increased radiosensitivity.
• Achieved a 12% higher TCP compared to standard equi-dose schedules in computational experiments.

Conclusions:

• An optimization method for non-uniform radiotherapy schedules was successfully developed.
• The method effectively models tumor response under hypoxia, leading to significant TCP improvements.
• This approach offers a promising strategy for enhancing radiotherapy efficacy.