RADT-23. AGE- AND LOCATION-DEPENDENT BENEFIT OF CHEMORADIOTHERAPY IN DIFFUSE MIDLINE GLIOMA: A POPULATION-BASED DISCOVERY AND SINGLE-CENTER VALIDATION

Abstract BACKGROUND Diffuse midline glioma (DMG) is a universally fatal CNS malignancy. Although radiotherapy is considered the standard of care, the added value of chemotherapy, particularly in the form of combined chemoradiotherapy (CRT), remains controversial, with current clinical guidelines offering conflicting recommendations across subgroups. METHODS We identified 409 patients with DMG from the SEER-22 database (2018–2021) and classified them into four treatment groups: no treatment, radiotherapy alone, chemotherapy alone, and combined chemoradiotherapy (CRT). Three-step multivariate Cox proportional hazards models were employed to assess the independent association between treatment modality and overall survival (OS). Subgroup analyses were stratified by age (<16 vs. ≥16 years) and tumor location (brainstem vs. non-brainstem). A four-quadrant stratification framework was developed to visualize survival trends across the intersection of age and tumor site. External validation was conducted using a real-world institutional cohort of 22 CRT-treated patients, with Kaplan–Meier survival analysis applied. RESULTS CRT significantly improved OS in multivariable models. However, survival benefits were predominantly observed in patients aged ≥16 or with non-brainstem tumors. Further research found demonstrated that only patients meeting both criteria - age ≥16 and non-brainstem location- achieved statistically significant survival improvement from CRT. These findings were replicated in the external validation cohort, reinforcing the robustness of the stratification model in real-world clinical settings despite the small sample size. CONCLUSION This study identifies a clinically actionable stratification paradigm for CRT use in DMG, based on a rare tumor population analyzed through population-based data and independently validated in a single-center cohort. Our results advocate for CRT in adult non-brainstem DMG, while highlighting the need for alternative approaches in pediatric or brainstem subgroups. These findings offer immediate implications for patient selection and clinical trial design in this high-mortality disease.