MEDB-51. Impact of residual tumor on outcomes in children and adolescents with medulloblastoma in the German HIT-cohort

INTRODUCTION: To date, a historical threshold of 1.5cm(2) extent of resection (EOR) of the primary tumor is used for risk stratification in pediatric medulloblastoma (MB). METHODS: Data of n=348 patients with MB included into the German HIT-MED studies and registries from 2000-2017 were retrospectively analyzed. Kaplan-Meier statistics and Cox regressions were calculated to determine the influence of EOR, histological and molecular subtype, MYC/N-amplification and presence of metastases at first diagnosis on overall and progression-free survival (OS/PFS). ROC curves were calculated aiming at identifying a new EOR cut-off. RESULTS: Median age at diagnosis was 6.8 [0.1-20.5] years. After initial surgery, residual tumor (R) ≥1.5cm(2) was present in 129/348 patients (37%). 57% presented with additional metastases. In most cases, classic histology was found (71%, DMB/MBEN: 20%, LC/AMB: 9%). Molecular subtype was nonWNT/nonSHH in 68% (Group 4: 37%, Group 3: 31%), WNT in 6% and SHH-activated types in 26% of evaluated cases (n=208). MYC/MYCN-amplification was present in 5 and 4%, respectively. 41/348 patients received additional surgery, resulting in 73% GTR (not reported: n=22). EOR had no significant impact on OS/PFS (GTR: 5y-OS/PFS 69.9±3.4/57.6±3.7%, STR: 5y-OS/PFS 71.4±4.1/56.2±4.5%, p=0.6/0.8; Cox: p=0.2/0.4; median follow-up 8.4 [0.1-18.0] years). ROC curves did not identify a significantly improved limit for outcome-relevant EOR cut-off. Analyses confirmed less favorable OS/PFS for patients with metastatic disease (p=0.04/p<0.05), LC/AMB (p<0.05/p<0.05), Group 3 (p<0.05/p<0.05), and MYC-amplification (p<0.05/p<0.05). Cox regression confirmed an increased risk for Group 3 (OS: HR=4.74, p<0.05, PFS: HR=2.90, p<0.05) and M+ (PFS: HR=2.69, p<0.05). DISCUSSION: No distinct influence of EOR </≥1.5cm(2) on OS/PFS in pediatric MB was detected. The relevance of EOR and the historical limit of 1.5cm(2) should be further investigated in distinct MB subtypes and in context of other risk factors.