Randomization of dose-reduced subcutaneous interleukin-2 (scIL2) in maintenance immunotherapy (IT) with anti-GD2 antibody dinutuximab beta (DB) long-term infusion (LTI) in front–line high-risk neuroblastoma patients: Early results from the HR-NBL1/SIOPEN trial.

10013 Background: We tested dose-reduced scIL2 in combination with DB-LTI and oral isotretinoin and evaluated toxicity and efficacy in high-risk neuroblastoma patients (EudraCT:2006-001489-17). Methods: High-risk patients (stage 4 ≥1y; stage 4 < 1y with MYCN amplification (MNA); stage 2, 3, 0-21y with MNA) received high intensity induction (rapid COJEC or N5-MSKC and TVD for insufficient response), surgery, high dose therapy with busulfan/melphalan and local radiotherapy. Patients ≤9 months between diagnosis and HDT/SCT who achieved at least a partial response prior to HDT/SCT and without progression thereafter were randomized to receive up to 5 cycles of 100mg/m2 DB-LTI (d8-17) ± 3x106 IU/m2 scIL2 (d1-5; d8, d10, d12, d14, d16) and 160mg/m2 oral isotretinoin (d19-32). Results: Between 04/2014 and 06/2018, 408 patients from 18 countries were randomized. Median follow-up is 1.8 years. Stage, age, MNA, induction treatments and remission status were well balanced between randomization arms. The 2yrs-EFS and -OS for DB-LTI (205 pts) vs. DB-LTI&scIL2 (203 pts) was 64%±4%vs63%±5% (p = 0.844) and 83%±3%vs74%±4% (p = 0.337). For patients in CR the 2yrs-EFS was 69%±5% for DB and 66%±6% for DB&scIL2. Patients with evaluable disease prior DB or DB&scIL2, the end of treatment response rate was 57% (26% CR, 31% PR) vs 52% (27% CR, 25% PR) with 2yrs-EFS rates of 58%±7% and 64% ±8%, respectively. Grade 3&4 toxicity was lower in the group with DB vs DB&scIL2 for fever (14%vs31%, p < 0.001) and pain (7%vs18%, p = 0.005), and no significant difference was seen for general condition (17%vs22%,ns), allergy (3%vs3%,ns), capillary leak (4%vs8%,ns), liver enzyme elevation (20%vs27%, ns) and neurological toxicities (2%vs2%,ns). Conclusions: We previously reported grade 3&4 toxicity to DB short-term infusion (STI) ± 10x6x106IU/m2 scIL2 for general condition (16%vs41%, p = 0.000), fever (14%vs40%, p = 0.000), allergic reaction (10%vs20%, p = p = 0.006), capillary leak (4%vs15%, p = 0.004), liver enzyme elevation (17%vs23%, ns), central neurotoxicity (3%vs8%, p = 0.034) and pain (16%vs26%, p = 0.048). Our results indicate that DB-LTI and dose-reduced scIL2 clearly reduced the toxicity profile, but showed absence of benefits of scIL2. DB-LTI achieved 2yrs-EFS in line with DB-STI (Ladenstein, Lancet Oncology 2018; Yu, NEJM, 2010) and a response rate > 50% supporting its use as standard of care IT. Clinical trial information: EudraCT:2006-001489-17.