Fakih Fleshes Out Perceptions on Pembrolizumab Approval in TMB-High Solid Tumors

The June 2020 FDA approval of pembrolizumab (Keytruda) for use in select patients with unresectable or metastatic solid tumors that are tumor mutational burden–high (TMB-H; ≥10 mutations/megabase) provides a potentially life-saving option to those who have exhausted all other available therapies, according to Marwan G. Fakih, MD, but additional biomarkers are needed to inform better patient selection.

The regulatory decision was based on data from a retrospective analysis of 10 cohorts of patients with previously treated, unresectable or metastatic, TMB-high solid tumors who were enrolled on the phase 2 KEYNOTE-158 trial.¹ Patients received pembrolizumab at 200 mg delivered intravenously every 3 weeks for 35 treatment cycles or until disease progression, unacceptable toxicity, or physician/patient choice.

Investigators determined that a total of 102 patients (13%) had tumors that could be identified as TMB-high. Results demonstrated that the overall response rate (ORR) with the agent in this group was 29% (95% CI, 21-39), and included a 4% complete response rate and a 25% partial response rate. Although the median duration of response (DOR) was not reached, 57% of patients experienced a response durations of ≥12 months and 50% of patients experienced response durations of ≥24 months.

“We're not necessarily going to see a response rate of 29% across all tumor types,” said Fakih, a professor in the Department of Medical Oncology & Therapeutics Research; medical director of the Judy and Bernard Briskin Center for Clinical Research; co-director of the Gastrointestinal Cancer Program; and section head of GI Medical Oncology at City of Hope. “That would be a super optimistic expectation, but these data do provide evidence that there is an increase in response rate [in those] with higher TMB.”

Despite these higher response rates, prior data have indicated a median overall survival (OS) of 11 months in the TMB-high group versus 13.3 months in the TMB-low group.² “One thing that does come up in this study is the median survival that has been reported; it could be similar for the TMB-high group and for the TMB-low group, despite the fact that the response rates were 29% versus 6% in favor of the TMB-high group,” said Fakih. “This tells us that additional biomarkers need to be discovered and validated.”

In an interview with OncLive, Fakih provided additional insight into the approval of pembrolizumab in patients with TMB-high solid tumors.

OncLive: Could you discuss the first tumor-agnostic FDA approval seen in 2017 with pembrolizumab?

Fakih: In 2017, the FDA approval of pembrolizumab was for patients with microsatellite instability and that was an important milestone because this was the first time that we saw an immunotherapy approved in a tumor agnostic fashion based on a biomarker.

What does the approval of pembrolizumab bring to patients with TMB-high solid tumors?

Recently, the FDA approved pembrolizumab for use in patients with a TMB of more than 10 or equal to or more than 10, irrespective of the primary tumor. Again, this is a tumor agnostic approval, and it was based on the KEYNOTE-158 trial. The prior approval for pembrolizumab in [patients with] microsatellite instability was also tumor agnostic and was based on several clinical trials that have evaluated pembrolizumab both in colorectal cancer (CRC), as well as in a tumor agnostic fashion based on microsatellite instability.

The KEYNOTE-158 clinical trial evaluated pembrolizumab in 10 rare malignancies; it wasn't open for all malignancies, and that may have been the cause for criticism by some as far as the implications of this study and then how it led to an approval in a tumor agnostic fashion. The tumor types that have been evaluated in the setting of the trial include anal cancer, biliary cancer, endometrial cancer, mesothelioma, neuroendocrine and salivary glands tumors, as well as small cell thyroid and bulbar cancer. All of these are relatively rare malignancies that have not been addressed by prior checkpoint inhibitor clinical trials.

The evaluation of the efficacy was based on predefined biomarkers, including TMB and PD-L1 expression. What's interesting is that the cut point of 10 and above for TMB was associated with an improved response rate in those patients. [Approximately] 29% of patients with a TMB of equal or more than 10 had an objective response, compared with 6% of patients who had a TMB of less than 10. Now, of course, this is not a randomized clinical trial and some clinical differences may exist in the patients with a TMB of less than 10 versus more than 10 on this study since those were not necessarily matched by other clinical criteria. In addition, some subgroups of patients had a higher prevalence of TMB of 10 and above than other subgroups of patients by tumor type.

Nonetheless, we are seeing a roughly three- to four-fold improvement in response rate in the TMB-high cohort. For the sake of this study, TMB-high was defined as 10 and higher. A substantial improvement in progression-free survival at 1 and 2 years [was observed] in the TMB-high cohort with pembrolizumab versus the lower TMB cohort with pembrolizumab. It suggests that this activity does transcend tumor type. Based on this study, the FDA has chosen to proceed with a tumor agnostic approval for pembrolizumab in any malignancy with a TMB of 10 or more.

In terms of safety, are the data consistent with what we previously have seen with pembrolizumab?

Yes, on the KEYNOTE-158 trial the safety was very much in line with what we typically see, and the incidence of severe adverse events (AEs) with the agent was [relatively] low. Of course, with any PD-1 inhibitor, you expect to see some cases of colitis. You definitely will see thyroiditis and hypothyroidism but, in general, the study shows that the treatment is quite well-tolerated, as has been shown from prior studies evaluating pembrolizumab.

Do you believe that there is sufficient evidence to warrant the approval?

We have to make a decision based on the data that we have. Certain malignancies are so rare that you will never be able to answer the question. If you set the bar high enough for every single primary tumor, as far as approval based on TMB, it will be very difficult and may take years to have those data and to conclusively answer that question. Of course, there is concern as to how you extrapolate these data, for example, with a patient with CRC who is microsatellite stable with a TMB of 11 or 12. Is this sufficient to tell us that those patients are going to benefit in a similar fashion?

Indeed, even within this study, as has been reported by the sponsor, there was also a higher response rate for the TMB above 13 in comparison with the TMB between 10 and 13. However, even within the subgroup of TMB of 10 to 13, the response rates were still double those of [patients who had a TMB of] less than 10. The fact that this is approved for patients who have exhausted the appropriate standard of care options, this does provide a lifeline for many of those patients who would otherwise be considered for hospice care or for a first-in-human phase 1 clinical trial.

Maybe this is an opportunity to start thinking about what we can do post-approval to better refine the treatment of those patients who are not included in these 10 disease types that were evaluated on KEYNOTE-158. I am comfortable, personally, offering pembrolizumab based on these data for a TMB of more than 10 in a patient who has exhausted standard-of-care options. However, it would be very helpful if we are able to generate a database to better understand in the future how beneficial this was for that particular cohort of patients and to better refine additional biomarkers for better patient selection.

Is there any other element of the trial or the approval itself that you wanted to highlight?

Not everyone will benefit from this treatment, even if their TMB is more than 10. The reality is that at the 2-year mark, perhaps 1 in 5 patients may still be without progression based on the curves that have previously been shown. That suggests that your median may not be that different because the median reflects the 50-percentile point. This tells us that, like we've seen in other tumor types such as gastric cancer, there is a subgroup of patients who are benefiting more than others, even within this TMB-high group.

However, I don’t believe that the lack of difference in the median overall survival (OS) is a suggestion that this is not an active therapy, given the fact of the differences in distribution of different subgroups or different types of cancers between the TMB-high and the TMB-low, and given the fact that the median does not represent the hazard ratio. I'm looking forward to the final data from this study that will better describe the OS and tell us more as to how much of an improvement in OS is noted at the 2-year and the 3-year mark and how many of the patients treated with this therapy derive a long-term benefit.

References

1. FDA approves pembrolizumab for adults and children with TMB-H solid tumors. News release. FDA. June 16, 2020. Accessed June 23, 2020. https://bit.ly/2NqAbbs.
2. Marabelle A, Fakih MG, Lopez J, et al. Association of tumour mutational burden with outcomes in patients with select advanced solid tumours treated with pembrolizumab in KEYNOTE-158. Ann Oncol. 2019;30(suppl 5):11920. doi:10.1093/annonc/mdz253.018