Phase I study of a novel glioblastoma radiation therapy schedule exploiting cell-state plasticity

Abstract

Background: Glioblastomas comprise heterogeneous cell populations with dynamic, bidirectional plasticity between treatment-resistant stem-like and treatment-sensitive differentiated states, with treatment influencing this process. However, current treatment protocols do not account for this plasticity. Previously, we generated a mathematical model based on preclinical experiments to describe this process and optimize a radiation therapy fractionation schedule that substantially increased survival relative to standard fractionation in a murine glioblastoma model.

Methods: We developed statistical models to predict the survival benefit of interventions to glioblastoma patients based on the corresponding survival benefit in the mouse model used in our preclinical study. We applied our mathematical model of glioblastoma radiation response to optimize a radiation therapy fractionation schedule for patients undergoing re-irradiation for glioblastoma and developed a first-in-human trial (NCT03557372) to assess the feasibility and safety of administering our schedule.

Results: Our statistical modeling predicted that the hazard ratio when comparing our novel radiation schedule with a standard schedule would be 0.74. Our mathematical modeling suggested that a practical, near-optimal schedule for re-irradiation of recurrent glioblastoma patients was 3.96 Gy × 7 (1 fraction/day) followed by 1.0 Gy × 9 (3 fractions/day). Our optimized schedule was successfully administered to 14/14 (100%) patients.

Conclusions: A novel radiation therapy schedule based on mathematical modeling of cell-state plasticity is feasible and safe to administer to glioblastoma patients.

Keywords: cell-state plasticity; clinical trial; glioblastoma; mathematical modeling; radiation oncology.

Conflict of interest statement

B.A. is an employee of Foundation Medicine and Roche. P.Y.W. receives research support from Astra Zeneca/Medimmune, Beigene, Black Diamond, Celgene, Chimerix, Eli Lily, ERASCA, Genentech/Roche, Kazia, MediciNova, Merck, Novartis, Nuvation Bio, Puma, Servier, Vascular Biogenics, and VBI Vaccines, and serves on the advisory board of Astra Zeneca, Bayer, Black Diamond, Boehringer Ingelheim, Boston Pharmaceuticals, Celularity, Chimerix, Day One Bio, Genenta, Glaxo Smith Kline, Karyopharm, Merck, Mundipharma, Novartis, Novocure, Nuvation Bio, Prelude Therapeutics, Sapience, Servier, Sagimet, Vascular Biogenics, VBI Vaccines. A.A. receives research funding from Varian and NH TherAGuIX and is consulting for Novartis and Seagen. F.M. is a cofounder of and has equity in Harbinger Health, has equity in Zephyr AI, and serves as a consultant for Harbinger Health, Zephyr AI, and Red Cell Partners. D.H.-K. is on the scientific Advisory Board of EmpNia, Inc., and Graegis’s Scientific. JD received research support from Novartis and is a consultant for Unum Therapeutics and Amgen. The authors declare that none of these relationships are directly or indirectly related to the content of this manuscript.

© The Author(s) 2022. Published by Oxford University Press on behalf of the Society for Neuro-Oncology.

Figures

Figure 1.

Mathematical modeling identifies a radiation therapy administration schedule predicted to improve survival of recurrent GBM patients. (A) Predicted tumor volume dynamics resulting from a radiation schedule of 51.6 Gy in 30 fractions with three times daily dosing with 3.25 h interfraction intervals compared to those from a schedule of 35 Gy in 10 fractions. (B) Predicted stem-like cell dynamics resulting from the administration schedules in (A). (C) Predicted effect of different two-phase radiation administration schedules on tumor volume dynamics. (D) Predicted stem-like cell dynamics resulting from the administration schedules in (C). QD—once daily dosing; TID—three times daily dosing.

Figure 2.

Patterns of tumor response and survival of patients treated in the clinical trial. (A) Kaplan-Meier plot for progression-free survival. (B) Kaplan-Meier plot for overall survival. (C) Cox proportional hazards regression model hazard ratios for progression-free survival. (D) Cox proportional hazards regression model hazard ratios for overall survival. (E) Swimmer plot of tumor response. The points indicate the responses at individual time points and the colors indicate the best response for each patient. (F) MRI scans of Patient 6 at the time of radiation therapy planning (left), 3 month follow-up (center) and 6 month follow-up with progression (right) showing a representative example of stable to regressed disease locally within the Planning Target Volume (contour line) but evidence of progressive enhancing, hypercellular disease with elevated cerebral blood volume immediately anterior to the Planning Target Volume. The patient died of disease 5 weeks after this time point. (G) MRI scans of Patient 1 at the time of radiotherapy planning (left) and 1 month follow-up (right) showing an example of a distant progression outside of the radiation therapy field (arrow), with stable to well controlled disease locally within the Planning Target Volume (contour line).