Dosimetric patterns of failure in the era of novel chemoradiotherapy in newly-diagnosed glioblastoma patients

Background: Patterns of failure (POF) may provide an alternative quantitative endpoint to overall survival for evaluation of novel chemoradiotherapy regimens with glioblastoma. Materials and Methods: POF of 109 newly-diagnosed glioblastoma patients per 2016 WHO classification who received conformal radiotherapy with concomitant and adjuvant temozolomide were reviewed. Seventy-five of those patients also received an investigational chemotherapy agent (everolimus, erlotinib, or vorinostat). Recurrence volumes were defined with MRI contrast enhancement. POF at protocol (POF_p), initial (POF_i), and RANO (POF_RANO) progression timepoints were characterized by the percentage of recurrence volume within the 95% dose region. POF_p, POF_i, and POF_RANO of each patient were categorized (central, non-central, or both). Results: POF of the temozolomide-only control cohort were unchanged (79% central, 12% non-central, and 9% both) across protocol, initial, and RANO progression timepoints. Unlike the temozolomide-only cohort, POF of the collective novel chemotherapy cohort appeared increasingly non-central when comparing POF_i with POF_p, with a non-central component increasing from 16% to 29% (p = 0.078). POF did not correlate with overall survival or time to progression. Conclusion: POF of patients receiving a novel chemotherapy appeared to be influenced by the timepoint of analysis and were increasingly non-central at protocol progression as compared with initial recurrence, suggesting that recurrence originates from the central region. Addition of everolimus and vorinostat appeared to influence POF, despite similar survival outcomes with the temozolomide-only control group. In studies dealing with novel therapeutic agents, robust and properly-timed dosimetric POF analysis may be helpful to evaluate biologic aspects of novel agents.