Background: BEACON-Neuroblastoma is a randomized Phase II trial to assess the activity of backbone chemotherapy regimens for children with RR-HRNB and to determine if inhibiting angiogenesis with bevacizumab adds to the activity of this chemotherapy. Methods: Patients aged 1-21 years with RR-HRNB with adequate organ function and performance status were randomized in a 2x2 factorial design to: temozolomide (T) versus irinotecan-temozolomide (IT), with or without bevacizumab. Here we report the results of the irinotecan randomization (T vs. IT), which had a probability-based Bayesian design. Primary endpoint was best overall response (complete or partial) during the first 6 courses, by RECIST for measurable disease patients and International Neuroblastoma Response Criteria for evaluable disease patients for which overall response rate (ORR) was calculated. Results: From 2013 to 2018, 61 patients were randomized to treatment with T and 60 to IT. Median age was 5.8 years. 85 and 36 had measurable and evaluable disease respectively; 55 and 66 had refractory and relapsed disease; 22 had MYCN amplification. Baseline characteristics were balanced between the arms. Response data was not yet available for 2 patients on T. Response was not assessable for 17 patients (did not have treatment or stopped early) who were considered non-responders. The ORR was 24% for T and 17% for IT (risk ratio (RR) = 0.70, 95% credible interval 0.32 to 1.44). The probability that the RR for ORR was >1.0 was 17%, meaning that IT did not show greater activity than T. There was no interaction between treatment with/without bevacizumab (heterogeneity test, p=0.7). 27 (44%) T and 35 (58%) IT patients had grade ≥3 toxicities as per CTCAE v4.0. Diarrhea occurred in no patients on T and 7 (12%) on IT; hematological toxicities included anemia (6 T, 4 IT), neutropenia (14 T, 22 IT) and thrombocytopenia (14 T, 11 IT). Conclusions: Irinotecan does not improve the response rate when added to temozolomide in RR-HRNB, but does increase diarrhea. Longer follow-up is needed before assessing whether it impacts progression-free or overall survival. Number of responses by treatment arm. Clinical trial information: NCT02308527