SURG-73. Predictors and Outcomes of Pathologically Determined Malignant Transformation in IDH-mutant Gliomas

INTRODUCTION The predictors and prognostic implications of malignant transformation (MT) in IDH-mutant gliomas remain poorly understood. Most studies include cohorts with mixed IDH status or define MT based on contrast enhancement, which is an unreliable marker of malignant progression. This study aims to investigate timing, predictors, and prognostic implications of MT in IDH-mutant gliomas. METHODS A retrospective review was conducted of patients with recurrent IDH-mutant gliomas who underwent multiple surgeries between 2009 and 2024 at the University of Pittsburgh Medical Center. MT was defined as a pathologically confirmed increase in tumor grade between the initial and any subsequent surgery. RESULTS A total of 108 patients were identified, of whom 64 (59.3%) experienced MT. Transformed cases included: G2–3 astrocytomas (n=20, 18.5%), G2–3 oligodendrogliomas (n=21, 19.4%), G3–4 astrocytomas (n=11, 10.2%), and G2–4 astrocytomas (n=12, 11.1%). Median time to transformation was 6.2 years (5.7–8.1) for G2–3 astrocytomas, 8.6 years (7–13.1) for G2–3 oligodendrogliomas, 5.1 years (4.8–NA) for G3–4 astrocytomas, and 4.3 years (4.1–NA) for G2–4 astrocytomas (p=0.51). Time to progression after subsequent surgery was shorter in MT vs. non-MT patients (1.1 years [0.8–1.4] vs. 4.4 years [1.9–NA], p<0.0001). Residual overall survival (OS)—from subsequent surgery to death or last follow-up—was also shorter in the MT group (1.7 years [1.4–3.8] vs. 5.8 years [2.9–NA], p<0.0001). However, OS from initial surgery was comparable (10.1 years [8.8–12.2] vs. 13.6 years [8.3–NA], p=0.2). Gross-total resection (HR: 0.40 [0.17–0.93], p=0.02), 1p/19q codeletion (0.36 [0.18–0.68], p=0.001), and adjuvant radiotherapy (0.28 [0.14–0.56], p=0.0003) were associated with decreased MT risk. CONCLUSION This study expands current knowledge on MT timing and impact in IDH-mutant gliomas. Future molecular studies are needed to clarify mechanisms and improve risk prediction.