Personalized prediction of tumor recurrence with image-guided physics-informed computational model in high-grade gliomas

High grade gliomas are infiltrating tumors characterized by their diffusive invasion and proliferative growth. Across and within patients heterogeneity of tumors makes it challenging to determine tumor spatial extent after surgical resection. Traditionally, tumor growth predictions after surgical resections rely on generalized models and population-based observations, which do not account for individual patient differences. To address this gap, we propose a personalized approach with image-guided computational model (digital twin) that incorporates physics-based modeling to predict tumor recurrence. Our digital twin involves an inverse modeling step, followed by a recurrence model that accounts for varying surgical effects. The physics-guided inverse model considers discrete loss, and estimates patient-specific diffusion (D) and proliferation (ρ) parameters from pre-operative magnetic resonance imaging (MRI) of 133 patients. The analysis is conducted using a publicly available dataset from The Cancer Imaging Archive (TCIA). The proposed model is personalized due to use of the patient-specific parameters gleaned from the real patient data to assess risk for both high-aggressive and low-aggressive tumor groups. The prognostic index for each patient reveals the interplay between tumor aggressiveness, surgical resection, and survival outcome. The results demonstrate that despite varying levels of surgical resections, patients with high-aggressive tumors have worse survival outcomes, with a median survival of 141-153 days due to rapid regrowth (0.10/day). In comparison, the low-aggressive group exhibits slower growth (0.06/day) and a median survival of 158-171 days. Furthermore, by integrating patient-specific diffusion and proliferation rates, the proposed method offers significant variability in tumor aggressiveness within high-grade gliomas.