BIOM-48. MOLECULAR AND RADIOLOGICAL MARKERS ASSOCIATED WITH MALIGNANT TRANSFORMATION IN IDH1 MUTANT LOW-GRADE GLIOMAS

Abstract BACKGROUND Malignant transformation (MT) of low-grade gliomas (LGGs) to a higher-grade variant seems inevitable. Therefore, identifying patients at risk of MT and establishing strategies to mitigate its development are essential. This retrospective study aims to assign predictive value to specific molecular and radiological markers associated with MT. METHODS Institutional database medical records of adult IDH1 mutant LGG patients were analyzed. MT was defined as histological confirmation of grade 3-4 after the 1st recurrence or new/increased post-progression enhancement (PPE) consistent with MT by multidisciplinary consensus. MT predictors were identified by logistic regression. The ki67 index optimal cutoff to predict MT was determined by calculating the maximal Youden index, employing receiver-operating characteristic (ROC) curves analysis to determine area under the curve (AUC). Survival was evaluated using Kaplan-Meier analysis and PPE’s predictive value for MT was calculated. RESULTS Among 53 patients with recurrent LGG, MT occurred in 33 (62.26%), majority being diagnosed with astrocytoma (61%), primarily in males (63.64%) with median age of 38 years, undergoing surveillance (60.61%) until first recurrence, molecular profile: high ki67 index (≥ 4 %) in 82%, p53 expressed in 40%, EGFR in 82%. Prognostic factors associated with MT were: high ki67 index (OR=2.75, 95% CI=2.89-21.32, p<0.05); 1p19q codeletion negative (OR=1.88, 95% CI=1.35-18.6, p<0.05); and EGFR expression (OR=1.95, 95% CI=2.99-17.85, p<0.05). At ROC analysis, PPE had a sensitivity of 0.69 and accuracy of 0.71 in predicting MT. A model combining high ki67 index and EGFR expression showed remarkable increase in both sensitivity (0.81) and accuracy (0.88). Survival analysis demonstrated higher post-recurrence and overall survival in 1p19q codeletion positive population (log rank p<0.05). CONCLUSIONS Immunohistochemistry profile heterogeneity was associated with MT. The proposed model exhibited superior potential compared to PPE in early prediction of MT. Further research is warranted to elucidate predictors and preventive strategies against MT occurrence.