Final Results from KEYNOTE-100 Show Promise for Pembrolizumab Monotherapy

Final results from the KEYNOTE-100 trial of pembrolizumab (Keytruda) in patients with advanced recurrent ovarian cancer, presented at the 2020 American Society of Clinical Oncology (ASCO) Virtual Scientific Program, indicated that the agent alone was associated with modest antitumor activity.

In an interview with CancerNetwork®, Ursula A. Matulonis, MD, of the Dana-Farber Cancer Institute, discussed the KEYNOTE-100 trial and what the final results mean for this patient population.

CancerNetwork®: Can you explain the study design for the KEYNOTE-100 trial?

Matulonis: So, these are the final results of the KEYNOTE-100 trial, and this is an international study that was started a number of years ago, examining the activity of the single agent pembrolizumab, so by itself. And obviously pembrolizumab is an anti-PD-1 checkpoint inhibitor for the treatment of women with advanced and recurrent ovarian cancer. So, the trial design was that patients were divided into 1 of 2 groups. Regardless of group, all patients had recurrent ovarian cancer, they had to have a good performance status of 0 or 1, they had to have also provision of an archival tumor tissue as well. So, for IHC, as well as other future biomarkers that are being done with this trial. The key exclusion criteria: our patients could not have mucinous histology. They could have any other histology, but could not have mucinous histology, no recent bowel obstruction, so no bowel obstruction within 3 months, and then no recent active autoimmune disease, and then no active central nervous system metastasis.

So, patients were enrolled into 2 cohorts, cohort A and then cohort B. Cohort A were less heavily pretreated patients, so they had to have at least 1 per line of treatment, and up to 3 prior lines of treatment. And the platinum-free or treatment-free interval was at least 3 months from the prior treatment and up to 12 months. Most of the patients went on this arm, about 285. And for cohort B, fewer patients, these patients in cohort B were more heavily pretreated. So, they had to have between 4 and 6 prior lines of treatment. And they had to have a prior platinum-free or treatment-free interval of at least 3 months. But there is no upper limit cap on the upper limits of the platinum-free or treatment-free interval. And those patients enrolled here were 91 patients. And importantly for the patients who are enrolled into cohort A, they represented, at least the first 100 patients into cohort A, they represented the training set. And these patients led to cutoff points for PD-L1 testing, so examining the CPS score, and basically giving us cutoff points where we would look at specific combined positive scores. Just to remind you, a combined positive score means that it's looking at the number of PD-L1 positive cells. So, it could be a cancer cell, could be a lymphocyte, could be a macrophage, a macrophage divided by the total number of tumor cells, times 100. And the cutoff points for CPS score were 1 or higher or 10 or higher.

So, what were the key takeaways of the final results?

So, the trial design was based upon overall response rate via RECIST 1.1 by blinded, independent central review either by cohorts, so cohort A versus B, and then also by CPS score, so 1 or higher 10 or higher. So, for the first primary endpoint, and I'm looking at overall response rate by cohort, the overall response rate for patients in cohort A was 8.1%, and for cohort B was 9.9%. So, for overall response rate for the 376 patients was 8.5. And the disease control rate was 1 of the secondary endpoints and that's the [complete response; CR] plus [partial response; PR] for stable disease more than 6 months. That was around 22% for both for both cohort A and cohort B.

Now, when we actually looked at the response rate, overall response rate based upon CPS score… So, for example, in patients who were in cohort A, less heavily pretreated patients had a CPS score of 1 or higher, responses were 6.9%; for CPS score 10 or higher, that response rate goes up to 11.6%. And then for cohort B, for CPS one or higher, overall response rate 10.2% and that goes up to 18.2% in patients for cohort B, in cohort B, who had a CPS score of 10 or higher. And if I just might point out that the higher CPS scores were more rare. So, in that patient population of cohort B, 10 or higher was only 18 patients. So, for all comers, all 376 patients for cohort A CPS score 1 or higher was 8%. And then for all the patient population, 376 patients, for CPS core 10 or higher, it was 13.8%.

Given these findings, what would you say are the overall implications of the study?

Yeah, so I think the key takeaway points are that pembrolizumab, like other single agent immune checkpoint inhibitors, in ovarian cancer has only moderate activity as a single agent. This has been observed with other single agent checkpoint inhibitors as well. The activity of pembrolizumab appears higher in ovarian cancers that have higher combined positive scores, so have higher levels of PD-L1 expression. And also, there was a higher trend towards increased response rate for patients with clear cell histology, and that's also been observed in other single agent checkpoint blockade studies in ovarian cancer. The median duration of response if a patient did have a response was around 10.2 months, which is very good. No new safety signals were identified. And further biomarker analysis is ongoing to try to really maybe potentially select out patients who may benefit more from single agent pembrolizumab.

Are there any sort of planned next steps for the trial?

So, I think the next steps for the study are the biomarker analysis. And this data that I presented at ASCO is the final data, so it is what it is. And I think the biomarker analysis may provide, you know, some clues about which patients may respond better than others. So, we'll have to see.

Is there anything that I didn’t ask you that you think is worth mentioning?

I think the other point I want to make is that we did not identify any new toxicities of pembrolizumab in this patient population who has recurrent ovarian cancer. So, most of the toxicities were around fatigue and nausea. The immune mediated toxicities were mostly thyroid gland, either hyperthyroidism or hypothyroidism. So, there was nothing new about the toxicities that was different than any other single agent pembrolizumab trial that's been done thus far.