SURG-09. Predicting Malignant Transformation in IDH-Mutant WHO Grade 2 Gliomas Using a Simplified Volume-Based Model

CNS WHO grade 2 diffuse gliomas with IDH mutations (IDHmut-LGGs) typically exhibit indolent growth but carry a substantial risk of malignant transformation (MT). In this study, we developed and validated a simplified mathematical model to estimate both the timing and likelihood of MT using only pre- and postoperative tumor volumes. This model represents a refinement of our prior approach, which relied on longitudinal MRI data and estimated cumulative tumor burden via trapezoidal integration. In contrast, the new model calculates cumulative tumor volume as the integral of an exponential growth function, enabling accurate MT risk estimation from minimal imaging data. We applied the model to a multi-institutional cohort of 269 patients with IDHmut-LGGs (118 oligodendrogliomas and 151 astrocytomas lacking contrast enhancement on MRI) from 10 centers across Japan. The MT risk predictions generated by the simplified model demonstrated high concordance with those of the previous serial-volume model, with Pearson correlation coefficients of 0.94 for oligodendrogliomas and 0.86 for astrocytomas, and significantly outperformed a time-based model (p < 0.01). Quantitative estimates revealed that, in oligodendrogliomas, the 5-year cumulative MT risk without surgical resection was approximately 41%, whereas reducing tumor volume to 5 cm³ postoperatively reduced the risk to 4.6%. Similarly, in astrocytomas, the corresponding risk decreased from 86% to 14% with comparable volume reduction. Tumor size–based stratification further revealed that lower pre- and postoperative tumor volumes are associated with delayed malignant transformation and improved overall survival (p < 0.05). To promote clinical applicability, we are currently developing a web-based decision-support tool that utilizes pre- and postoperative tumor volumes to estimate MT risk and timing. This platform will enable clinicians to simulate the impact of surgical resection extent, thereby supporting data-driven surgical planning and individualized patient care. Our model offers a practical and robust framework for improving outcome prediction and treatment strategies in IDHmut-LGGs.