Pembrolizumab for the treatment of disease relapse after allogeneic hematopoietic stem cell transplantation

A failed graft-versus-tumor (GVT) effect is a common mechanism of relapse after allogeneic hematopoietic cell transplantation (alloHCT). Although targeting the PD-1/PD-L1 axis may restore GVT effects, PD-1 blockade exacerbates graft-versus-host disease (GVHD) in murine models, and severe GVHD can occur in patients treated with anti-PD-1 therapy after alloHCT. Therefore, we developed a prospective study to assess the safety and efficacy of pembrolizumab in patients relapsing after alloHCT. Eligible patients received pembrolizumab (200 mg every 3 weeks) for up to 2 years. Twelve patients were enrolled (8 patients with acute myeloid leukemia, 1 patient with myelodysplastic syndrome, 1 patient with classical Hodgkin lymphoma, and 2 patients with diffuse large B-cell lymphoma [DLBCL]). All participants received reduced-intensity preparative regimens with in vivo T-cell depletion. The median time from alloHCT to enrollment was 587 days (range, 101-4211). Three participants (25%) experienced grade 3 to 4 immune-related adverse events (irAE) (pneumonitis, 2 patients; hyperthyroidism, 1 patient), all occurring after 1 to 2 cycles, and resolving after pembrolizumab discontinuation and corticosteroid treatment. irAEs of any grade occurred in 5 patients (42%). No treatment-emergent GVHD was observed. Overall and complete response (CR) rates were 22% (2/9). Both patients achieving CRs had PD-L1 gene–amplified lymphomas and diffuse PD-L1 expression on pretreatment biopsies. An acquired EZH2 mutation was identified at relapse in a patient with DLBCL who achieved an initial CR to pembrolizumab, which was associated with downregulated HLA expression on malignant B cells, implicating EZH2 mutations as a potential immune escape mechanism after PD-1–blockade therapy. In conclusion, after alloHCT, treatment with pembrolizumab is feasible and associated with objective responses in relapsed lymphoid malignancies but can induce severe irAEs, requiring vigilant monitoring. This trial was registered at www.clinicaltrials.gov as #NCT02981914.