A Robust Optimization Approach to Cancer Treatment under Toxicity Uncertainty

Area of Science:

• Oncology
• Mathematical Optimization
• Pharmacology

Background:

• Optimal treatment protocols are crucial in cancer therapy but sensitive to patient-specific parameter variations.
• Parameter uncertainty in drug effects and patient attributes challenges clinical application of optimal protocols.

Purpose of the Study:

• To develop a robust optimization framework for cancer treatment protocols considering parameter uncertainty.
• To model and mitigate toxicity uncertainty, specifically low absolute neutrophil count (ANC), in tyrosine kinase inhibitor (TKI) therapy for chronic myeloid leukemia (CML).

Main Methods:

• Formulation of treatment protocol design as a robust optimization problem (ROP).
• Application of a mixed integer ROP to minimize cumulative tumor size.
• Modeling toxicity uncertainty, focusing on the variable rate of ANC decrease within a defined interval.

Main Results:

• The developed optimization methods identified dosing schedules that significantly reduce tumor size.
• The proposed protocols demonstrated effectiveness in preventing tumor recurrence over 360 weeks.
• Toxicity constraints, particularly concerning ANC levels, were satisfied across all uncertain parameter realizations.

Conclusions:

• Robust optimization provides a viable approach to designing cancer treatment protocols resilient to parameter variations.
• The study successfully demonstrated the efficacy of ROP in optimizing TKI therapy for CML, balancing tumor reduction with toxicity management.
• The findings highlight the importance of considering parameter uncertainty in developing effective and safe cancer treatment strategies.