P12.32.A MOLECULAR STRATIFICATION BY PTEN STATUS REVEALS DIFFERENTIAL SECOND-LINE THERAPY RESPONSES IN RECURRENT IDH-WILDTYPE GLIOBLASTOMA

Abstract BACKGROUND Recurrent IDH-wildtype glioblastoma (rGBM) has a poor prognosis, and second-line therapies (i.e. nitrosoureas, regorafenib, bevacizumab) offer limited benefit. No clinical or molecular factors have been validated as predictors to second-line therapy. PTEN is frequently mutated in these tumors, influencing neurogenesis and malignancy, but his predictive role in therapy response remains unclear, as histopathologic grading often fails to capture the underlying molecular complexity. MATERIAL AND METHODS We conducted a retrospective, two-institution study (Veneto Institute of Oncology, Padua and University of Turin, Italy) of consecutive rGBM patients treated between 2019 and 2022. We determined PTEN status by FoundationOne®CDx and TSO 500 illumina; PTEN with a pathogenic mutation or deletion was considered altered. We evaluated outcomes using RANO criteria and survival analysis adjusted for age, ECOG PS, second surgery, steroid use and MGMT promoter methylation. This analysis accounted for known prognostic factors using Kaplan-Meier and Cox regression, isolating PTEN’s specific impact RESULTS A total of 227 patients (pts) were enrolled (regorafenib n=95; bevacizumab n=39; nitrosoureas n=92). The mOS from the date of treatment was: 7.3 ms (95% CI 6.4-9) with Nitrosoureas (NS), 12 ms (95% CI 9.1-14) with Regorafenib (Reg) and 10 ms (95% CI 6.8-13) with Bevacizumab (Bev). PTEN was altered in 58 pts (61%) treated with Reg, 8 pts (57%) with Bev and 23 pts (58%) with NS. In univariate analysis PTEN alteration was associated with short survival in Reg treated (mOS of 9.4 ms VS 17 ms in alterated VS wild-type (wt) PTEN, respectively; HR 2.04 (95% CI 1.26-3.30), p=0.043) and NS treated cohort (median OS of 6.9 ms VS 8.0 ms in alterated VS PTENwt, respectively; HR 1.63 (95% CI 1.06 -2.51), p=0.027). In contrast, in Bev treated cohort PTEN alteration did not reach a statistical significance in univariate analysis (mOS 7.2 ms versus 10 ms in alterated VS PTENwt, respectively; HR 1.23 (95% CI 0.58-2.60) p=0.6). Of note, in multivariate analysis PTEN alteration maintained a significant impact as predictor of short survival after treatment for both Reg and NS cohort (HR 1.88, p=0.016 and HR= 1.97, p=0.020, respectively) together with MGMT status (HR 0.42, p<0.01 and HR=0.54, p=0.023, respectively). CONCLUSION Pathogenic PTEN alteration may be a predictor of Reg and NS efficacy in rGBM patients. However, a prospective study with a larger population is needed to better define the role of PTEN in these patient populations.