ONCO-35

Determining the impact of tumor heterogeneity on radiation dose planning via MRI-based mathematical modeling. Tarini Thiagarajan 1, Thomas E. Yankeelov 1-6, Bikash Panthi 7, Caroline Chung 7, David A. Hormuth II 1-2. 1 Oden Institute for Computational Engineering and Sciences,  2 Livestrong Cancer Institutes,  3 Biomedical Engineering,  4 Diagnostic Medicine, and  5 Oncology, The University of Texas at Austin.  6 Department of Imaging Physics,  7 Radiation Oncology, MD Anderson Cancer Center.

SMB2025 SMB2025 Follow
Share this

TariniThiagarajan

Oden Institute for Computational Engineering and Sciences
"Determining the impact of tumor heterogeneity on radiation dose planning via MRI-based mathematical modeling. Tarini Thiagarajan 1, Thomas E. Yankeelov 1-6, Bikash Panthi 7, Caroline Chung 7, David A. Hormuth II 1-2. 1 Oden Institute for Computational Engineering and Sciences,  2 Livestrong Cancer Institutes,  3 Biomedical Engineering,  4 Diagnostic Medicine, and  5 Oncology, The University of Texas at Austin.  6 Department of Imaging Physics,  7 Radiation Oncology, MD Anderson Cancer Center."
Radiotherapy planning typically assumes homogeneous efficacy across the tumor, which can lead to an overestimation of tumor cell death and control. We seek to quantify these errors by identifying the radiation dose boost required by non-homogeneous treatment efficacy models to yield the tumor response predicted by homogeneous response. We collected data from ten glioblastoma patients who received a total of 60 Gray (Gy) of radiation delivered in 30 fractions, concurrent chemotherapy, and were imaged prior to, and then at one-, three-, and five-months post-therapy. These data were used to inform a patient-specific, mechanically coupled reaction-diffusion model describing the spatiotemporal progression of tumor growth and response to therapy. The radiotherapy effect was modeled as an instantaneous decrease in the tumor volume fraction (ϕ) after each dose, with the surviving fraction (SF) defined as the ratio between the post- and pre-treatment ϕs. For each patient, the proliferation rates, diffusion coefficients, and SFs were calibrated to data up to five months post-therapy. We then simulated tumor growth using the calibrated model with (a) homogeneous SF, or heterogeneous SF, based on (b) vascularity or (c) cell density. We determined the patient-specific global dose boost required for models (b) and (c) to match the response predicted by model (a) one-month post-radiotherapy for SFs of 0.2-1.0 in increments of 0.05. For 0.55-0.95 SFs, model (b)’s predicted response required an additional 0.67 +/- 0.30 (mean +/- standard deviation) Gy per day, while model (c) only needed an additional 0.27 +/- 0.18 Gy per day across all patients. There was a statistically significant (P < 0.05, Wilcoxon rank sum test) difference between dose predictions for models (b) and (c) across all SFs. We developed an approach to calculate the radiation dose increases needed by non-homogeneous treatment efficacy models to match the tumor response predicted by a homogeneous model.
Additional authors:


SMB2025
#SMB2025 Follow
Annual Meeting for the Society for Mathematical Biology, 2025.