10047-CO-1 A Predictive Model of Malignant Transformation in IDH-Mutant LGG to Support Clinical Decision-Making

Abstract IDH-mutant lower-grade gliomas (IDHmut-LGGs) generally exhibit relatively slow progression. However, a significant proportion of cases undergo malignant transformation (MT) within several years, which greatly impacts overall prognosis. We previously developed a mathematical model to estimate the risk and timing of MT based on tumor volume calculated from all available MRI scans during the clinical course (Aoki et al., Cancer Res, 2021). While this model was useful for retrospective analysis, it required longitudinal imaging data and thus had limited applicability in prospective clinical practice. In this study, we developed a simplified mathematical model that dynamically estimates MT risk and timing without the need for serial imaging. The model uses only pre- and postoperative tumor volumes and treatment information as inputs, and assumes exponential tumor growth with treatment-specific growth coefficients. We applied this model to 269 patients with IDHmut-LGGs (118 oligodendrogliomas and 151 astrocytomas) from 10 institutions across Japan. The predicted values showed strong concordance with those of the previous model, with correlation coefficients of 0.94 for oligodendrogliomas and 0.86 for astrocytomas. Furthermore, the model significantly outperformed a simple time-based model (p < 0.01). Using this model, we quantitatively assessed the effect of surgical tumor volume reduction on MT risk in cases with an initial tumor volume of 50 cm3. In oligodendrogliomas, the 5-year MT risk was estimated at 38.52% without resection, which decreased to 9.47% when the postoperative volume was reduced to 10 cm3, and to 6.36% when reduced to 5 cm3. In astrocytomas, the corresponding MT risk was 85.70% without surgery, which decreased to 25.23% and 14.17% with postoperative volumes of 10 cm3 and 5 cm3, respectively. We are currently developing a web-based clinical decision-support tool based on this model, which is expected to assist in prospective treatment planning and optimization of surgical strategies in routine clinical settings.