Abstract BACKGROUND Low-grade glioma (LGG) is the most common childhood central nervous system tumor. Carboplatin/vincristine (CV) is a standard of care regimen used for incompletely resected tumors. Monthly carboplatin is an alternative regimen that has not been prospectively compared to CV previously in untreated patients. METHODS Patients ≤ 21 years with LGG and no treatment besides surgery were eligible. Patients were stratified according to NF1 status and randomized to Arm A (carboplatin/vincristine) or Arm B (carboplatin monthly). The primary endpoint was 3-year progression-free survival (PFS). Secondary endpoint analysis is underway and includes quality of life, tumor response rates, toxicity and association of molecular alterations on outcomes. Analyses included log-rank test and descriptive statistics. RESULTS There were 95 evaluable patients with a median age of 5 (range 0-18) years; 67 non-NF1 subjects (32 Arm A, 35 Arm B) and 28 NF1 subjects (14 Arm A, 14 Arm B). Three-year PFS was superior for patients without NF1 treated on Arm A (58%; 95% CI: 38%, 74%) versus Arm B (33%; 95% CI: 18%, 49%) (p=0.04). In patients with NF1, there was no statistical difference in PFS between those treated on Arm A (71%; 95% CI: 41%, 88%) and Arm B (93%; 95% CI: 59%, 99%) (p=0.08). Grade 3/4 toxicity neutropenia was more commonly seen in Arm A patients compared to Arm B for both NF and non-NF1 patients. Peripheral neuropathy was also more commonly seen in non-NF1 patients (Grade 1-4; 32%) compared to NF1 patients (7%) for patients treated on Arm A. There were 15 carboplatin reactions (Grade 1-4); 7 non-NF1 on Arm A, 3 non-NF1 on Arm B, and 5 NF1 on Arm A. CONCLUSIONS Non-NF1 patients treated with carboplatin and vincristine show statistically significant improved PFS compared to monthly carboplatin. Patients with NF1 are statistically underpowered due to low accrual.