ASCO 2018 | What happened with pembrolizumab in MM clinical trials?

Aviva Krauss, a hematologist/oncologist from the US Food and Drug Administration (FDA) was a member of the review committee that called a halt to several clinical trials assessing the use of pembrolizumab to treat multiple myeloma (MM) patients, in combination with commonly used backbone regimens. Speaking from the 2018 American Society of Clinical Oncology (ASCO) Annual Meeting on 1 June, Dr Krauss provided an overview of the FDA analysis, providing greater clarity with regards to why the trials were halted and what the future might be for the use of pembrolizumab to treat MM. In particular, the immune-related adverse events (irAEs) and their association with response in two clinical trials were detailed.

Dr Kraus explained that there has been an explosion of clinical trials using PD-1 or PD-L1 inhibitors with more than 1,500 clinical trials testing 50 clinical stage agents targeting PD-1 or PDL1, and that this figure rises even further taking into account combination trials. However, most of these trials target solid tumors and only pembrolizumab (pembro) and nivolumab have been approved for use in hematological malignancies, specifically for relapsed and refractory classical Hodgkin’s Lymphoma. The rationale for moving forward with these agents in MM was sound, as the tumor microenvironment (TME) plays a big role and myeloma cells express PD-L1, whereas normal plasma cells do not. However, initial trials with pembro monotherapy failed in MM, so studies moved forward with standard combination regimens. Two such studies were Keynote-183 and Keynote-185, which were halted by the FDA in July 2017 - see previous MM Hub article detailing the decision to pause the trials, which were later called to a full halt.

Key Data:

All data is given as pembrolizumab arm vs control-arm:

Keynote 183:

• Pembro in combination with pomalidomide and dexamethasone to treat relapsed and refractory (RR) MM patients (pts)
• Cut-off: 2 June 2017
• Number of patients (pts) safety population: 120 vs 121
• Median follow-up = 8.1 months
• Deaths (pts): 29 vs 21
• Cause of death included: myocarditis, infections, cardiac, respiratory, Stevens-Johnson syndrome, and unknown causes
• Overall survival (OS) hazard ratio (HR) = 1.61 (95% CI: 0.91, 2.85)
• ORR = 34% vs 40%

Keynote-185:

• Pembrolizumab plus lenalidomide and dexamethasone in newly diagnosed (ND) MM pts ineligible for transplant
• Number of patients (pts) safety population: 149 vs 145
• Median follow-up = 6.6 months
• Deaths = 19 vs 9
• OS HR = 2.06 (95% CI: 0.93, 4.55)
• ORR = 64% vs 62%
• No difference was observed in either trial in overall response rate or time-to-progression between arms

To evaluate the survival data, the assessment team set out to better understand the effects of irAEs, a well-documented effect of these inhibitors. IrAEs have been widely reported and can affect any organ, but whether there is an association between the occurrence of irAEs and an increased rate of response is a topic of debate, as there is data to support either scenario. Therefore, Dr Krauss and her team used data from the two keynote trials to address this question. Due to the subjective definition of irAE, specific criteria for a range of side-effects was defined ahead of the data analysis.

• Keynote-183: Patients with irAE = 58% vs 45%; G≥3 irAE = 18% vs 13%
• Keynote-185: Patients with irAE = 68% vs 44%; G≥3 irAE = 36% vs 8%

The development of irAEs did not significantly change ORR in the pembro arm, but development of irAE was associated with increased ORR in both arms. When they compared patients that had no AEs with patients that did, there was no significant difference in response rate between the two categories. However, when they selected patients that presented with grade 3 irAE or greater, the differences were more pronounced, and in particular in NDMM patients. The reason for this difference could be attributed to drug exposure and how healthy the immune system is. Patients that have relapsed MM will have a more exhaustive immune system, whereas newly diagnosed patients will, in general, have a more robust immune system. However, in this particular case, the benefit of pembro in NDMM patients was outweighed by the increased toxicity.

Conclusion

Overall, a decreased OS was observed in two randomized trials using pembrolizumab in both NDMM and RRMM patients. In the RRMM population, there was no difference in ORR with or without irAE, whereas in the NDMM population increased ORR was observed in patients with irAE and increased irAE rate. Finally, Dr Krauss cautioned against the use of single-arm trials, explaining that “phase II trials can provide us with exponentially more critical safety information than that which can be gleaned with single-arm trials”.