Mathematical Optimization Approaches to Radiation Therapy Treatments of Brain Metastases

Radiotherapy is a common form of cancer treatment for brain tumors, with the goal of irradiating tumors while minimizing damage to adjacent healthy tissue. Treatment plans are unique to each patient and are computed using large-scale optimization algorithms. Treatments are usually delivered over several consecutive days, allowing healthy cells to recover between sessions; this concept is known as fractionation. Today, treatment plans typically adhere to a uniform fractionation approach, delivering identical doses of radiation each day. We study an alternative treatment approach called non-uniform fractionation for patients with multiple metastatic lesions in their brain. For these patients, irradiating different subsets of lesions on each day may reduce harm to healthy brain tissue. We propose a new formulation and computational approach to computing non-uniformly fractionated plans and quantifying the differences in healthy brain tissue damage between uniform and non-uniform fractionation.