Pembrolizumab plus lenalidomide and dexamethasone for the treatment of multiple myeloma: results from KEYNOTE-183, KEYNOTE-185 and KEYNOTE-023

In solid tumors, PD-1/PD-L1 pathway inhibitors have advanced treatment options, however, their application in multiple myeloma (MM) is unclear.¹ In a pre-clinical in vitro study by Güllü Topal Görgün and colleagues, anti-PD1 and anti-PD-L1 agents, alone, or in combination with lenalidomide, induced effector cell-mediated anti-myeloma cytotoxicity in samples from patients with relapsed/refractory MM (RRMM).² However, studies using PD-1/PD-L1 inhibitors as monotherapy have shown little benefit in tumor regression.^1,3

Based on the preclinical results, and early clinical studies using single-agent PD-1 inhibition, it was hypothesized that using PD-1/PD-L1 inhibitors in combination with other agents might prove to be more efficacious in MM. Studies were, therefore, initiated investigating PD-1/PD-L1 inhibitors in combination with immunomodulatory agents (IMiDs), based on their synergistic mechanism of action.¹

Pembrolizumab is a highly selective, humanized monoclonal antibody that targets the PD-1 checkpoint and has exhibited anti-tumor activity. The phase III KEYNOTE studies; KEYNOTE-183⁴ and KEYNOTE-185⁵, and the phase I study, KEYNOTE-023⁶ are examples of studies investigating the use of pembrolizumab in combination with pomalidomide (P) or lenalidomide (R) and dexamethasone (d). (Table 1).

Table 1. Comparison of key features of pembrolizumab-based KEYNOTE studies^4-6

On July 3 2017, the United States (US) Food & Drug Administration (FDA) halted all three KEYNOTE trials due the increased risk of using pembrolizumab, compared to the potential benefit, based on interim analysis of the phase III trials.⁷ This article summarizes these study findings and the potential future applications of PD-1 checkpoint inhibitors in MM.

KEYNOTE-183 and KEYNOTE-185^4,5

At the request of the FDA, an unplanned interim analysis of KEYNOTE-183 and KEYNOTE-185 studies was conducted, both of which were published in Lancet Hematology in July 2019. The results of KEYNOTE-183 were published by Maria-Victoria Mateos, University Hospital of Salamanca, Salamanca, ES, and colleagues⁴ and the results of the interim analysis of KEYNOTE-185 were published by Saad Zafar Usmani, Levine Cancer Institute/Atrium Health, Charlotte, US, and colleagues.⁵

Study design and patient characteristics

Both KEYNOTE-183 and KEYNOTE-185 combined pembrolizumab with dexamethasone and an IMiD, KEYNOTE-183 used pomalidomide, and KEYNOTE-185 used lenalidomide. KEYNOTE-183 was conducted in patients with RRMM (N= 249) who had an Eastern cooperative oncology group (ECOG) score of one or two and had previously received ≥ two prior lines of therapy, excluding pomalidomide.⁴ KEYNOTE-185 enrolled patients with NDMM (N= 301) who were ineligible for autologous stem cell transplantation (ASCT) with an ECOG score of zero or one and who were treatment naïve.⁵

In both studies, 200mg of pembrolizumab was given intravenously (IV) every three weeks, in combination with 4mg of pomalidomide or 25mg of lenalidomide, orally, on days 1–21 and 40 mg of dexamethasone, weekly (reduced to 20mg for patients over the age of 75). Treatment was continued until disease progression (PD), unacceptable adverse events (AEs) or withdrawal from study.^4,5

Table 2. Patient characteristics in KEYNOTE-183 and KEYNOTE-185^4,5

KEYNOTE-183 had co-primary endpoints of progression-free survival (PFS) and overall survival (OS).⁴ In KEYNOTE-185, the primary endpoint was PFS which was investigator-assessed as the trial was terminated early.⁵ The response rate was assessed using International Myeloma Working Group (IMWG) 2011 response criteria in both studies.⁸ Efficacy analysis was conducted in all randomly assigned patients, with safety analysis performed in patients receiving ≥ 1 dose of pembrolizumab. The median follow-up for KEYNOTE-183 was 8.1 months (interquartile range [IQR]: 4.5–10.9) and for KEYNOTE-185 was 6.6 months (IQR: 3.4-9.6).^4,5

Adherence to study

Table 3. Discontinuations in KEYNOTE-183 and KEYNOTE-185 with reasons

Discontinuations related to safety and AEs are shown in Table 5.

KEYNOTE-183:

• Median duration of treatment: 123.5 days (57.5–225) vs 127 days (78–253)
• Median number of cycles: 4.4 (2–9)

KEYNOTE-185:

• Median duration of treatment: 131 days (IQR: 44.0–253.0) vs 162 days (IQR: 83–246)
• Median number of cycles: 5.0 (IQR: 2–9) vs0 (IQR: 3–9)

Efficacy

• KEYNOTE-183: Median PFS was shorter in the pembrolizumab-Pd arm compared to the Pd arm, indicating pembrolizumab increased the risk of progression or death.
• KEYNOTE-185: There were more deaths in the pembrolizumab-Rd arm compared to the Rd arm giving an unfavorable HR, again indicating a higher risk of death with pembrolizumab. In total, 19 patients died in the pembrolizumab-Rd arm (six due to PD and 13 due to AEs) and nine in the Rd arm (one due to PD and eight due to AEs).
• Due to the early termination of the studies, efficacy analyses were underpowered, however the interim results are shown in Table 4.

Table 4. Efficacy analysis of KEYNOTE-183 and KEYNOTE-185

Safety

KEYNOTE-183: Adverse events (AEs) of any grade were reported in 99% of patients in the pembrolizumab-Pd arm and 96% in the Pd arm (Table 5). AEs with a difference of ≥ 5% between arms were; neutropenia, anemia, fatigue, pneumonia, nausea, headache, back pain and increased alanine aminotransferase levels. Grade 3-4 AEs with a difference of ≥ 5% between arms (pembrolizumab-Pd vs Pd) were neutropenia (34% vs 21%) and thrombocytopenia (12% vs 7%).

KEYNOTE-185: The safety analysis is shown in Table 5. There were more grade 3-5 events in the pembrolizumab-Rd group than the Rd group. No grade 5 events with over a 5% difference between both arms occurred.

Table 5. Safety analysis of KEYNOTE-183 and KEYNOTE-185

Table 6. Disease characteristics amongst patients who died during KEYNOTE-183 and KEYNOTE-185

Conclusions of KEYNOTE-183 and KEYNOTE-185

Both studies, despite incomplete data and a shorter treatment exposure than anticipated, found pembrolizumab has an unfavorable risk-benefit profile in patients with RRMM and NDMM. Authors of KEYNOTE-183 hypothesized that the lack of improvement in outcomes may have been due to the immunodeficiency seen in the patient population who were heavily pretreated. Meanwhile, the authors of KEYNOTE-185 found that despite more deaths due to AEs in the pembrolizumab arm, no specific AE was attributable to this increase, and suggested that an imbalance of risk factors between arms contributed to early progression, subsequent deaths and TRAEs.

KEYNOTE-023⁶

The KEYNOTE-023 study was a phase I study which investigated the use of pembrolizumab-Rd in patients with RRMM, as opposed to NDMM as in the phase III KEYNOTE-185 study.

This study aimed to identify the maximum tolerated dose (MTD) of pembrolizumab in combination with Rd. At the time the FDA halted the trials, the median follow-up was 18.9 months (0.8–36.0). The MTD of pembrolizumab was determined to be 2mg/kg, which was subsequently administered as 200mg fixed dose. The optimal doses of R and d in combination with pembrolizumab were determined to be 25mg and 40mg, respectively.

In total, 66 patients were enrolled, and 62 were treated; 33.9% of patients had previously received over five lines of therapy and 33.9% had ISS stage III disease. Over the course of the trial, 47 patients (75.8%) discontinued treatment, mainly due to PD (54.8%).

All patients experienced an AE. Two deaths were attributed to AEs; one due to veno-occlusive liver disease that was unrelated to study treatment, and one due to ischemic stroke related to lenalidomide. TRAEs occurred in 59 patients (95.2%), with 37 (59.7%) being grade 3-5.

The ORR was 44% with a 94% disease control rate and 50% of patient had stable disease. The median PFS was 7.2 months (95% CI, 3.9–12.3) and median OS was NR. Estimated survival at 2-years was 64.7%.

Future perspectives of PD-1 inhibitors in MM

PD-L1 has been linked to disease progression in patients with monoclonal gammopathy of undetermined significance (MGUS) and smoldering MM (SMM), and is specifically and highly expressed on malignant plasma cells, making it an attractive therapeutic target. Patients with RRMM have high PD-L1 concentrations, hypothesized to play a role in clonal resistance. However, the use of PD-1 and PD-L1 inhibitors has not shown activity in single-agent studies, leading to the KEYNOTE studies of combination approaches.⁹

All three KEYNOTE studies in this article have been stopped by the FDA, and as such, no long-term efficacy data or quality-of-life assessments will be collected. Whilst this will not change current clinical practice , it has provided invaluable data that will be used for future trial design.^5,6 Other studies involving checkpoint inhibitors in MM (NCT02726581, NCT01592370, and NCT02612779), were also stopped by the FDA.⁹

In KEYNOTE-183, the authors noted the association between the disease stage and immune system dysfunction, stating that PD-1 blockade is more likely to be safer and effective in patients with a lower disease burden and healthier immune systems.⁴ Additionally, despite randomization, more patients who died prematurely had high-risk features at baseline such as ISS stage III disease and high-risk cytogenetics, in both KEYNOTE-183 and KEYNOTE-185.^4,5

In the KEYNOTE-185 study, since no single AE was attributed to the higher risk of death in the pembrolizumab arm, the authors suggested future studies using PD-1 inhibitors in combination with other agents are justified. However, stricter patient selection should be considered, along with the other factors shown in Table 7.^4,5

The efficacy and safety results of the phase Ib KEYNOTE-023 were promising in patients with RRMM. However, the study was not permitted to continue due to the FDA hold, based on the lack of benefit observed in the phase III KEYNOTE-185 and KEYNOTE-183 studies.⁶

Table 7. Considerations for future clinical trials of checkpoint inhibitors in MM^4,5,9

If the optimal combination is identified, it will be important for clinicians to be alert to the toxicity profiles and immune-related AEs that are novel to this new treatment.⁹