Randomized Phase II Trial of Dendritic Cell/AML Fusion Cell Vaccination Compared to Standard of Care Therapy in AML CR1

Introduction: We have developed a personalized cancer vaccine in which patient derived AML cells are fused with autologous dendritic cells (DCs), presenting a broad array of antigens including shared and neoantigens. Here we report the results of the multicenter randomized phase II clinical trial (NCT03059485) of DC/AML fusion cell vaccination in AML patients who achieve a remission following chemotherapy or HMA/Venetoclax.

Methods: Eligible patients were >55 years and were not candidates for allogeneic transplantation. The primary endpoint was to assess 2-year progression free survival. Leukemia cells were collected and cryopreserved at time of disease presentation. Patients underwent standard of care remission induction therapy, with either 7+3 or HMA/Ven, at the discretion of the treating physician. Patients in CR (CR/CRi) were randomized 2:1 to the vaccine (Arm A) or standard of care arm (Arm C). Arm B, vaccine with PD-1 blockade, was removed prior to any patient being treated. Patients in Arm C could receive maintenance therapy, while those in Arm A did not. Patients in Arm A underwent leukapheresis and DCs were generated by cytokine mediated differentiation. DC/AML fusions were created by co-culture of DCs and leukemia cells in the presence of polyethylene glycol. The trial was designed with a target accrual of 75 patients but closed to prematurely due to slow enrollment during COVID-19 and changes in standard of care leukemia therapy including maintenance HMA/Ven and IDH/FLT3 inhibitors. Herein, we report the results of 41 randomized patients. This trial is not powered to detect difference in clinical endpoints.

Results: 41 patients with median age 68 years (range 55-86) were randomized, 27 to Arm A and 14 to Arm C. Two patients (one in each arm) withdrew consent and were not followed for response or survival. Additionally, In Arm A, 1 patient had disease progression, 1 proceeded to allogeneic transplantation, 1 had autoimmune hepatitis and were considered ineligible to proceed with vaccine generation. Fusion vaccine was successfully generated in 23 patients, with a mean fusion efficiency of 50.73%, and a viability of 75%.

2-year PFS in arm A was 31% by intention to treat and 36% in vaccinated patients (N=22) compared to 23% in arm C (N=13). 2-year OS was 73% in Arm A and 61% in Arm C. 2 grade-2 injection site reactions, 2 grade-2 and 1 grade-3 platelet count decrease were attributable to vaccine.

Immune response was characterized by interferon gamma (IFN) secretion following ex-vivo exposure to autologous tumor lysate. A mean of 1.5-fold increase in the IFN+ CD4+ T cells and 1.85-fold in the IFN+ CD8+ T cells was observed following vaccination (pre-vaccine 1 compared to peak post-vaccination, n=19). Deep TCR repertoire sequencing of serial blood and bone marrow samples was performed from pre- and post-vaccination timepoints for four patients. These longitudinal data were used to identify T cell clonotypes that significantly increased following vaccination. Two major classes of treatment-expanded clonotypes were identified: pre-existing clonotypes that increased following vaccination and emergent clonotypes only detected following vaccination. Comparison of 3- and 6-month follow up blood samples demonstrated persistence of all emergent clones.

Conclusion: In summary, DC/AML fusion vaccination in older patients in CR1 in the absence of maintenance therapy resulted in a 2-year PFS of 36% and OS of 73%. Immunologic response was characterized by emergence and persistence of specific T cell clonotypes. Ongoing studies include analyzing expanded T cell clones as a means for antigen discovery and evaluation of DC/AML vaccination following allogeneic transplant. Future directions include the potential to develop a novel combinatorial approach to maintenance therapy, by incorporating DC/AML fusion cell vaccination to HMA and targeted therapy based maintenance therapy for AML.