Background: Neuroblastoma (NBL) is the most common extra-cranial solid tumor in children and outcomes for children with relapsed/refractory (RR) NBL remain dismal. While the combination of chemoimmunotherapy (CIT) with the anti-GD2 antibody dinutuximab as per COG protocol ANBL1221 has improved responses, this regimen is still not curative. Dinutuximab can recruit NK cells for anticancer activity, however, NK cells are depleted in NBL patients undergoing chemotherapy. We devised a method to expand allogeneic NK cells ex vivo to improve effector function and render the NK cells resistant to TGFβ-induced suppression (TGFβi). These TGFβi NK cells are expanded from a universal donor (UD) pool, enabling multiple cycles of “off-the-shelf” high-dose therapy. We hypothesize that adoptive transfer of UD-TGFβi NK cells to patients with RR NBL sequentially following CIT is safe and improves outcomes compared to CIT alone. Methods: This Phase I/II study evaluates the safety and tolerability of UD-TGFβi NK cells in combination with CIT and compares anti-cancer efficacy to CIT alone. Eligibility includes age < 30 years, RR or progressive NBL, and prior treatment with at least 4 cycles of induction chemotherapy. The treatment protocol consists of 21-day cycles of CIT as per COG protocol ANBL1221 with UD-TGFβi NK cells (1x10⁸ cells/kg) infused on day 8. Up to six cycles of treatment are permitted. Overall response rates are defined using the Revised International NBL Response Criteria, with 95% confidence intervals compared to the results from ANBL1221. Progression-free and overall survival will be estimated by the Kaplan Meier method. Results: Four subjects (7 – 11 years) have been enrolled. All patients presented with high risk NBL: three with relapsed disease and one with refractory disease. All received at least one prior salvage therapy. Three patients have completed all 6 cycles of protocol therapy, and one remains on treatment. 20 NK cell infusions have been administered to date. All NK cell infusions occurred within the protocol specified window. One subject experienced grade 1 fever with two NK cell infusions. There have been no other adverse events attributable to the NK cells. Adverse events attributable to CIT were similar to those described in COG ANBL 1221. Treatment delays for cytopenias due to CIT occurred in two patients. Three patients achieved a partial response. A fourth patient attained prolonged stable disease (9 months) but experienced a skeletal recurrence. Conclusions: UD allo-TGFβi NK cells can be safely and feasibly administered to children with RR NBL after treatment with CIT, with early objective responses observed in the preliminary cohort of patients. Ongoing studies will assess markers of response, NK cell persistence, pharmacokinetics and pharmacodynamics. Clinical trial information: NCT04211675.