Differentiating tumor recurrence from radiation-induced necrosis: An image-based mathematical modeling framework

Patients with intracranial metastases often undergo stereotactic radiosurgery (SRS) for local control. Following SRS, some patients develop radiation-induced necrosis, which appears radiographically similar to tumor recurrence on follow-up imaging. Both may appear as an enhancing lesion in MR T 1 -weighted contrast enhanced imaging with surrounding FLAIR abnormality, complicating diagnostic and therapeutic efforts. We develop a spatiotemporal model of tumor growth in this work to parameterize tumor growth kinetics, based on contrast enhanced Ti-weighted serial MR imaging. In a proof-of-concept study to demonstrate feasibility of the framework, we evaluated two patients, one with recurrence and one with radiation-induced necrosis. Model-data fits were used to parameterize tumor cell diffusion coefficient and tumor cell proliferation rate. Differences between the pathologies were found when comparing the tumor cell proliferation rate, suggesting the potential of this model to distinguish between diagnoses in a biophysical model-based image analysis framework.