Pembrolizumab for the Treatment of Disease Relapse Following Allogeneic Hematopoietic Cell Transplantation

Background: Disease relapse remains the primary cause of mortality following allogeneic hematopoietic cell transplantation (alloHCT). One important mechanism of disease relapse in this setting is failure of the graft-versus-tumor (GvT) effect, and the PD-1/PD-L1 axis may diminish GvT after alloSCT. We hypothesized that PD-1/PD-L1 interactions prevent donor-derived T cells from eliminating malignant cells expressing minor histocompatibility antigens, and that blocking PD-1/PD-L1 interactions with the anti-PD-1 antibody, pembrolizumab (pem), might restore GvT and induce clinical responses in patients (pts) with relapsed hematologic malignancies following alloHCT. However, PD-1 blockade therapy has been associated with severe graft-versus-host disease (GVHD) in murine models, and GVHD has been reported in humans treated with anti-PD-1 therapy after alloHCT. Thus, we developed a prospective clinical study to test the tolerability and preliminary efficacy of pembrolizumab in patients with relapsed leukemia/lymphoma after alloSCT.

Methods: Pts with AML, MDS, or B cell lymphomas with biopsy-proven recurrence after alloSCT were eligible, as long as no active acute GVHD > grade 1 or chronic GVHD was present. Pts were treated with pem 200 mg IV q3 weeks for up to 2 years, provided that neither intolerable side-effects nor disease progression occurred. Pem could be delayed for treatment-limiting toxicities (TLT), defined as immune-related adverse events (irAEs) not meeting criteria for a dose-limiting toxicity (DLT). DLT was defined as the development of grade 3 or 4 acute GVHD/irAE, any unexpected grade > 2 toxicity related to pem, or development of > grade 2 vital organ dysfunction secondary to an irAE within 90 days of pem initiation. A two-stage mini-max design was chosen, with an early stopping rule for DLT after the first 11 patients were enrolled.

Results: 11 pts (7 male, 4 female), mean age 49.5 yrs (range, 27-62 yrs) have been enrolled. 8 pts had AML and 3 had lymphoma (DLBCL - 2, cHL - 1). 6 pts had matched-related donors (MRD) and 5 pts had haploidentical/umbilical cord blood (haplo-cord) donors. Pts with MRD were conditioned with fludarabine, melphalan, and alemtuzumab, or fludarabine and busulfan. Pts with haplo-cord donors were conditioned with fludarabine, melphalan, and ATG. 5 pts had prior acute GVHD. Pts relapsed following alloHCT at a median of 453 days (range, 101-1021 days). A median of 2 cycles of pembrolizumab (range, 1-8) was administered. 3 pts are receiving ongoing treatment. 3 pts experienced a DLT due to an irAE (grade 3-4 pneumonitis 2 pts; grade 3 hyperthyroidism 1 pt), all of which occurred after 1-2 cycles of pem, and resolved after pem discontinuation and corticosteroid treatment. 1 pt experienced a TLT (grade 2 rash), but resumed pem treatment. Among all pts, irAEs of any grade occurred in 7 pts. 7 pts were evaluable for response. 3 pts (2 AML, 1 DLBCL) experienced progressive disease (PD), 2 pts (AML) had stable disease (SD), and 2 pts achieved CR (DLBCL, cHL). 1 pt with AML (myeloid sarcoma) in whom pem was discontinued for PD by PET/CT imaging had a concurrent tumor biopsy that revealed marked T cell infiltration and PD-L1 expression on a significant fraction of malignant myeloid cells, suggestive of possible inflammatory "pseudo-progression". 1 pt in CR developed therapy-related AML unrelated to pem. Notably, both patients with CR following pem had PD-L1 gene-amplified lymphomas by FISH, and diffuse PD-L1 protein expression on pre-treatment biopsies. Currently, 4 pts have died, all due to disease progression, and 7 are alive. A total of 26 patients are expected to be enrolled.

Conclusions: Treatment with pem in the post-alloHCT disease relapse setting is feasible, but can induce early and severe irAEs, requiring vigilant monitoring. To date, objective responses were seen in 2/3 lymphoma patients treated with pem. In AML, pem may be less effective, where a best response of SD was observed in 2 pts, and possible "pseudo-progression" in a patient with myeloid sarcoma. This study continues to accrue pts, and correlative analyses are underway. To our knowledge, these are the first prospective data of PD-1 blockade therapy in the post-alloHCT setting.