Randomized use of anti-GD2 antibody dinutuximab beta (DB) long-term infusion with and without subcutaneous interleukin-2 (scIL-2) in high-risk neuroblastoma patients with relapsed and refractory disease: Results from the SIOPEN LTI-trial.

10014 Background: We determined the role of scIL-2 combined with long term infusion (LTI) of DB in patients (pts) with high-risk relapsed/refractory neuroblastoma. Methods: 160 pts were enrolled into an open label SIOPEN Phase II clinical trial (EudraCT 2009-018077-31). Pts were randomly assigned to receive up to 5 cycles of 100 mg/m 2 DB-LTI (d8-17) and 160 mg/m 2 oral isotretinoin (d19-32) (81 pts) with and without 6x10 6 IU/m 2 scIL-2 (d1-5; 8-12) (79 pts). Endpoints were toxicity, response rates and 2yrs-event free and -overall survival. Results: Between 07/2014 and 07/2017, 160 pts from 11 countries were randomised. Median follow-up is 2.6 years. Pts were well balanced between arms according to stage, age, MYCN amplification, patients with relapse and remission status. The 2yrs-EFS and -OS for DB (81 pts) vs. DB combined with scIL-2 (79 pts) was 59%±6% vs 65%±6% (p = 0.721) and 79%±5% vs 84%±4% (p = 0.904). In 97 pts with evaluable disease, a response rate of 49% (9% CR, 40% PR) vs 52% (26% CR, 26% PR) after treatment with DB vs DB and scIL-2 was observed. Grade 3&4 fever (16% vs 46%, P = 0.000), allergic reaction (1% vs 14%, P = 0.004), hematological toxicity (46% vs 66%, P = 0.013) and neurotoxicity (0% vs 8%, p = 0.003) were significantly worse in the combination arm, but no difference was seen for capillary leak, gastrointestinal, liver enzyme elevation and pain. Paraplegia possibly related to the treatment was observed in 2 pts in the combination arm, none in the arm without scIL-2, and one resolved to baseline. A subgroup of 34 pts who had a relapse and measurable disease at treatment start, showed a 2yrs-EFS and -OS in DB (17 pts) vs DB combined with scIL-2 (17 pts) of 35%±12% vs 69%±12% (p = 0.116) and 59%±12% vs 81%±10% (p = 0.167). However, this trend was statistically not significant. Pharmacokinetic and HACA response between both arms was not different with overlapping antibody concentration-time curves and a HACA response of 15/81 (19%) (DB) vs 16/79 (20%) (DB and scIL-2). Conclusions: No significant difference in efficacy of DB combined with scIL-2 and increased toxicity in this arm suggests that this schedule of scIL-2 is of no additional benefit. Clinical trial information: 2009-018077-31.