RADT-79. PATTERNS OF HIGH-GRADE GLIOMA RECURRENCE AFTER MRI-GUIDED ADAPTIVE RADIOTHERAPY

Abstract Following standard-of-care radiotherapy using baseline imaging for target definition prior to start of radiation, high-grade glioma (HGG) often recurs within the high dose region or marginally. MRI-guided adaptive radiotherapy (ART) has emerged as a strategy to refine treatment targeting by adjusting plans in response to anatomical changes during treatment. We retrospectively analyzed recurrence patterns in 50 patients with histologically confirmed HGG enrolled in a prospective study of weekly multiparametric MRI (mpMRI) during chemoradiation. Radiation treatment was adapted at the clinical discretion of the treating physician based on MRI changes. Plan adaptation occurred at variable time points during the treatment course. Fourteen cases (28%) were adapted during the treatment, while 36 cases were not adapted. Recurrence was classified as in-field, marginal, or out-of-field relative to the original 95% isodose volume. Follow-up MRIs were rigidly co-registered to the planning CT in RayStation (v11B). A recurrent gross tumor volume was delineated when recurrence overlapped the CTV. The cases requiring treatment plan adaptation had greater gross tumor volumes at baseline (median 49.1 [IQR: 25.5–92.9] vs, 27.5 [IQR: 24.0–55.1] cm³). Of 14 cases that were adapted, 11 recurred (10 in-field, 1 out-of-field), with no marginal failures. Of 36 cases non-adapted, 30 recurred: 27 in-field, 2 marginally, and 8 out-of-field, with some patients exhibiting overlap across categories. IDH-wildtype status was significantly associated with in-field recurrence (p = 0.0015), while MGMT status and extent of resection were not associated with recurrence location. Based on these initial results of this prospective study, ART was associated with fewer marginal recurrences, despite larger tumors at baseline. These findings highlight the feasibility and potential benefit of ART in HGG. Further investigation of the molecular and advanced imaging features in the recurrent tumors may reveal approaches for improved tumor targeting to reduce marginal recurrence and potential dose-escalation for in-field resistance.