DIFFERENTIAL TREATMENT EFFECTS OF STANDARD AND HYPOFRACTIONATED RADIATION REGIMENS IN GLIOBLASTOMA PATIENTS

The identification of the subgroups with differential treatment effects (DTE) is important for decisionmaking in personalized treatment. The DTE analysis assists in identifying patients who are more likely to benefit from a particular treatment regimen. The aim of the study was to analyze DTE in terms of the survival of glioblastoma (GBM) patients in the groups of standard radiotherapy (SRT) and hypofractionated radiotherapy (HRT) by the multicluster modeling of homogenous groups while retaining the statistical characteristics of the overall primary study cohort.

The cohort of 159 patients with newly diagnosed GBM stratified according to the radiotherapy regimen (HRT group (n = 110/69.2%); SRT group (n = 49/30.8%)) was evaluated retrospectively. Forty-eight subgroups (multiclusters) were created by enumerating all possible combinations of 5 significant covariates (age, sex, the radicality of the surgical resection, chemotherapy, and Karnofsky performance status) of the Cox model. The DTE for the cancerspecific survival (CSS) within 48 modeled multiclusters were studied by comparing the interpolated Weibull CSS curves according to the Kolmogorov - Smirnov test.

The findings showed that the SRT group was superior to the HRT group by CSS only in 3 of the modeled clusters presenting clinical scenarios with a non-radical tumor resection, no chemotherapy, and low Karnofsky functional status (≤ 70 scores) (Cluster 10: male aged < 60; Cluster 21: female aged ≥ 60; Cluster 22: male aged ≥ 60). Most of the studied clinical variants (45 of 48 multiclusters) did not demonstrate a significant difference when comparing the interpolated Weibull curves of the CSS for the SRT and HRT groups according to the Kolmogorov - Smirnov test (p ≥ 0.05).

We propose a novel multicluster modeling approach that addresses DTE in relatively small samples of GBM patients receiving SRT or HRT. This original analytical method can be taken into consideration while designing new well-powered prospective trials aimed at the subgroup analysis in GBM patients who will be most beneficial from personalized treatment strategies.