Background: While the addition of anti-GD2 immunotherapy led to improvement in outcomes in patients on the Children’s Oncology Group (COG) ANBL0032 study, relapse remains a concern. Prior studies demonstrated the prognostic importance of time to first relapse, however, the effect of immunotherapy on timing and patterns of relapse in neuroblastoma (NBL) have yet to be evaluated. The purpose of this exploratory analysis was to describe the impact of immunotherapy on patterns of relapse in patients with high-risk NBL, including a descriptive comparison of sites of relapse based on post-consolidation treatment received [dinutuximab with cytokines and isotretinoin (DIN) vs isotretinoin alone (ISO)]. Methods: A retrospective, descriptive analysis of patients on ANBL0032 was performed, including patients randomized to DIN or ISO and those non-randomly assigned to DIN after ISO arm closure. Pt characteristics including age, stage, MYCN amplification status, tumor grade, mitosis-karyorrhexis index (MKI) and ploidy were summarized descriptively and relapse sites were tabulated. For DIN patients who subsequently relapsed, overall survival (OS) was calculated starting from the time of first relapse after enrollment on ANBL0032 (“post-relapse OS”). Kaplan-Meier OS curves were generated based on site of relapse. Results: The analytic cohort included 1,431 (DIN = 1,327; ISO = 104) patients. Among DIN patients, 492 relapsed, many in > 1 site. In the randomized cohort (n = 248), 122 relapsed (DIN = 68/144; ISO = 54/104). The frequencies (DIN; ISO) by site of relapse in the randomized cohort were: bone (53%; 54%), CNS (16%; 11%), lymph node (13%; 17%), abdominal (10%; 17%), paraspinal (6%; 2%), liver (3%; 4%), other soft tissue (22%; 7%). A higher proportion of ISO patients had marrow relapse (29.4% DIN; 48.2% ISO); however, the proportion of DIN patients with lung relapses appeared higher (9% vs 2%). Among all relapsed patients, the proportion with bone relapse did not appear to differ between treatment groups, regardless of MYCN status. Among patients with MYCN amplified disease, the proportion with marrow relapse did not appear to differ based on treatment [21/149 (14.1%) DIN; 3/20 (15.0%) ISO]; however, among patients with MYCN non-amplified disease, the proportion with marrow relapse appeared higher in the ISO group [16/23; 69.6%] vs the DIN group [52/193 (26.9%)]. Conclusions: In this exploratory analysis of patients on COG ANBL0032, the pattern for site of relapse appears to differ between patients treated with DIN vs ISO. While immunotherapy remains the treatment of choice in this population, the findings from this retrospective exploratory analysis warrant further investigation to decrease the risk for post-immunotherapy relapse. Clinical trial information: NCT00026312.