2260 - Association of Changes in Circulating Immune Cells With Survival in GBM Patients Treated With Chemoradiation
Presenter(s)
S. Workman1, A. Li1, Y. Tarui2, S. Pokala3, I. C. Liu1, L. G. Ross4, L. Sloan5, K. J. Redmond6, R. Agyekum1, and L. R. Kleinberg1; 1Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, 2Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, 3Department of Molecular and Cellular Biology, Johns Hopkins University, Baltimore, MD, 4Department of Radiology, Johns Hopkins University, Baltimore, MD, 5University of Minnesota: Department of Radiation Oncology, Minneapolis, MN, 6Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University, Baltimore, MD
Purpose/Objective(s): Circulating immune cell populations, including lymphocytes, monocytes, and neutrophils, have been associated with antitumor immune responses. We aimed to develop a composite immune score using standard blood metrics to characterize immune state before and during chemoradiation and assess associations with survival in glioblastoma (GBM).
Materials/Methods: We retrospectively reviewed 314 GBM (WHO 2007 CNS classification) patients treated with conventionally fractionated radiotherapy (2010–2020) with available white blood counts at baseline, 2 weeks, and 6–12 weeks post-treatment. Immune scores were calculated as z-scored composites standardized to baseline values (zlymphocytes – 0.25zmonocytes – 1.25zneutrophils), with coefficients predicated upon maximization of model likelihood for survival within biologically plausible range. Cox regression, repeated measures ANOVA, and Levene’s test were used to evaluate associations with overall survival (OS) and longitudinal variation. Immune recovery was defined as the change between 6–12 week and post-treatment scores.
Results: Median age was 59 years; median OS was 65 weeks. Median immune scores were 0.152 (baseline), 0.039 (post-treatment), and 0.122 (6–12 weeks). Repeated measures ANOVA indicated no significant differences over time (F(2,626) = 0.68, p = 0.51), but post-treatment variance decreased (Levene’s test = 20.05, p < 1×10?8). Baseline and 6–12 week immune scores were protective on univariate Cox regression (HR 0.81, p = 0.0008; HR 0.66, p < 1×10?¹°, respectively), whereas post-treatment scores were not prognostic (HR 0.89, p = 0.057). Immune recovery was associated with OS univariately (HR 0.79, p < 0.005) but not after adjusting for 6–12 week immune score (HR 0.96, p = 0.54). In multivariate models including age, MGMT, and IDH status, 6–12 week immune score remained significant (HR 0.74, p < 0.005; concordance = 0.63), as did baseline score in a separate model (HR 0.84, p = 0.01; concordance = 0.60).
Conclusion: Composite immune scores prior to and 6–12 weeks after chemoradiation are robust predictors of survival in GBM. Absolute 6–12 week immune status dominates prognostic value over early recovery, highlighting the importance of post-treatment immune reconstitution. Although mean immune scores remained stable over time, post-treatment variance was significantly reduced, suggesting transient compression of systemic immune heterogeneity. Prognostic discrimination re-emerged at 6–12 weeks, indicating that immune reconstitution and restoration of inter-patient immune variability underlies the survival association.