A Phase I/Ib Study of Nivolumab for Relapsed Hematologic Malignancies after Allogeneic Hematopoietic Cell Transplantation (alloHCT)

BACKGROUND: Patients (pts) with relapsed hematologic malignancies (HM) after alloHCT are a unique population, given the potential to harness a dormant graft-vs.-tumor effect therapeutically. We previously reported that CTLA-4 blockade with ipilimumab was feasible and active in this population (Davids et al., N Eng J Med, 2016). In retrospective studies, anti-PD1 antibodies were active in pts with relapsed lymphoid malignancies after alloHCT, though with substantial toxcity due to GVHD (Herbaux et al. and Haverkos et al., Blood, 2017). Here, we report on the first prospective clinical trial of PD1 blockade in pts with relapsed HM after alloHCT.

METHODS: The primary objectives in this ph I/Ib, multicenter, investigator-initiated, CTEP-sponsored study (CTEP 9204) were to determine MTD and evaluate safety of nivolumab (nivo). Secondary objectives were to assess efficacy and immunologic correlates. Pts with any HM with relapse or persistent disease after alloHCT were eligible. Nivo was initially given to a 1 mg/kg cohort, with planned escalation to a 3 mg/kg cohort or de-escalation to a 0.5 mg/kg cohort depending on toxicities. Nivo was dosed q2 wks until progression or unacceptable toxicity, and disease-specific response evaluations were q4 cycles.

RESULTS: A total of 28 pts with relapsed HM after alloHCT were treated. Median age was 57 (range 27-76), and pts had the following HM: AML (n=11), MDS (n=7), Hodgkin lymphoma (HL, n=5), non-Hodgkin lymphoma (NHL, n=3), MPD and CLL (n=1 each). Median number of prior therapies was 2 (range 1-9), and 18/28 (64%) had progressed after at least 1 prior therapy for relapse post alloHCT. The median time from alloHCT to study enrollment was 21 mo. (range 5.7-174 mo.).

Six pts were treated initially with nivo 1 mg/kg. Two immune-related adverse events (irAEs) resulted in DLTs, including one pt with sepsis and fatal ARDS, and one pt with new anti-phospholipid antibodies and a fatal thrombotic cerebral vascular accident. Other irAEs included gr3 pneumonitis and transaminitis (n=1 each). One pt had cGVHD (NIH mild). Response was observed in 3/6 pts, including 1 CR (PMBCL) and 2 PR (HL and CMML).

Due to the toxicities at 1 mg/kg, a cohort of 8 pts was then treated with nivo 0.5 mg/kg, which was generally well-tolerated, with no DLTs. A phase Ib expansion cohort then accrued 14 more pts at 0.5 mg/kg. Accrual was terminated after 14 pts were treated due to meeting the protocol-defined stopping rule of ≥4 DLTs in the first 15 pts in this cohort. These DLTs included 2 cases of grade III acute GVHD (liver and gut) as well as gr3 elevated bilirubin (n=1) and gr3 transaminitis (n=1) which did not recover to ≤gr1 within 4 wks. The 2 pts with liver dysfunction without histological evidence of GVHD eventually improved, but both pts with GVHD died due to complications from GVHD. Other toxicities included gr4 lipase elevation, gr3 rash, gr3 transaminitis, gr3 orthostatic hypotension, and gr2 seizure in a pt with a known seizure disorder (n=1 each). In the 22 pts treated at nivo 0.5 mg/kg, 10 pts (45%) had new onset or worsening of GVHD, including 1 with aGVHD only, 7 with cGVHD only (3 of whom had baseline cGVHD), and 2 with both acute and cGVHD. Shorter time from alloHCT was significantly associated with higher risk of developing GVHD (p=0.019). The overall response rate in the 19 evaluable pts treated at nivo 0.5 mg/kg was 16%, including 1 pt with HL with CR and 1 pt each with HL and AML achieving PR. Nine pts had stable disease for at least 1 response evaluation, and 7 pts had progressive disease as best response.

Studywide, the overall response rate was 24% (6/25), the median number of cycles received was 3 (range 1-25), and 12/28 (43%) had at least 1 dose delay due to toxicity. With a median follow-up of 3.9 mo. (range 1.4-20.9 mo.), the 6 mo. PFS and OS were 39% and 61%, respectively.

CONCLUSIONS: In this first prospective clinical trial of an anti-PD1 antibody for relapsed HM post-alloHCT, severe GVHD and irAEs occurred, even at the lower dose of nivo 0.5 mg/kg, leading to early closure due to toxicity. Modest anti-tumor activity was observed mainly in lymphoid malignancies known already to be responsive to anti-PD1 therapy, which may justify further exploration of anti-PD1 therapy in those populations in trials with strategies to mitigate toxicity; however, given the more favorable safety and efficacy profile of anti-CTLA-4 therapy in other HM, our future studies focus on combining ipilimumab with novel partners to improve outcomes.