Patient Responses to I-O in CSCC

Shubham Pant, MD:I think that’s a very important question. It’s important to distinguish between PD-L1 [programmed death-ligand 1] staining, tumor mutational burden, and microsatellite instability [MSI]. Sometimes it’s kind of confusing for people. I know you’ve talked about it, but can you parse them out, those 3? What is the PD-L1 staining? What is a tumor mutational burden, per se? And what are microsatellite unstable tumors? How are they different?

Nikhil Khushalani, MD:Sure. So PD-L1 staining is primarily identifying the intensity and the proportion of cells, both tumor cells as well as in lymphocytes, in the actual tumor biopsy specimen. So that is a numerical value that....

Shubham Pant, MD:Like immunohistochemistry, you stain it and you look at how much you see, basically.

Nikhil Khushalani, MD:Correct. The mutational burden is detected primarily by next-generation sequencing. That is the actual mutational load or the number of mutations seen within a coding segment of the genome. So, again, the tumors that have high mutational burden, such as melanoma, CSCC, MSI-high cancers, those tend to respond better. Contrast that with certain pediatric tumors or low-grade tumors, and even uveal melanoma, the other disease that I treat. They really don’t have a high degree of mutations at all. So several tests that are directly on tumor tissue identify mutational burden.

Microsatellite instability, I think you’re more of the expert than I am. Starting with colorectal cancer, specifically looking at microsatellite stability/instability genes, and from an immunohistochemical standpoint, can be done. Or it can also be done on next-generation sequencing. CD8 density, that’s primarily looking at the tumor itself and looking through immunofluorescence assays, understanding what the percentage of CD8-positive cells right at the tumor interface is.

And several publications have now highlighted, there’s such a vast difference between tumors in this regard. Some tumors, such as melanoma, it’s almost exclusively on the tumor cell, the PD-L1 staining. In some tumors, it’s literally at the interface of the tumor and the lymphocytes. And in some gastric cancers, we actually see them more on the lymphocytes alone and not on the tumor at all. I think we’re still trying to scratch the surface, and maybe we’re going down the wrong biomarker. I don’t know the answer to that and time will tell us.

Shubham Pant, MD:And maybe the different biomarkers for different disease.

Nikhil Khushalani, MD:Absolutely.

Shubham Pant, MD:So, for immune checkpoint inhibitors, CSCC could have a different biomarker than gastric, lung, and such.

Nikhil Khushalani, MD:Absolutely.

Shubham Pant, MD:Regarding response in melanoma, you said it was higher in treatment-naïve patients. What were the response rates seen with immune checkpoint inhibitors in CSCC?

Nikhil Khushalani, MD:In the EMPOWER study, and this was done in 2 parts, as I mentioned earlier. You had a phase I part and a phase II part. And what was reported in theNew England Journal of Medicinepublication was a compilation from both of those parts. In the phase I part, it was 26 patients. In the phase II part, which was reported specifically for metastatic CSCC, the locally advanced cohort had not yet reached maturity in terms of data. So when one collates this together, we had a total of 85 patients on that clinical trial at the time of reporting. The vast majority were male, most of them with good functional status. And, as we mentioned, many of them had already been heavily pretreated either with chemotherapy and/or radiotherapy.

Shubham Pant, MD:I think some were in their 90s, right?

Nikhil Khushalani, MD:That is correct. If you look at both modalities, yes, you’re absolutely correct. And, of course, surgery had been part of their armamentarium that had already been received previously. In terms of when one parses it out in terms of the phase I portion, the response rates were approximately 50%. So half of the patients treated with cemiplimab….

Shubham Pant, MD:Had shrinkage of the tumor.

Nikhil Khushalani, MD:Had shrinkage of the tumor as defined by RECIST [response evaluation criteria in solid tumors] 1.1 criteria, which is greater than 30% decrease in the sum of the largest dimensions for the measurable tumors. And that’s quite remarkable; 50% actual response rate and about a 20% stable disease rate. And, as you know very well with solid tumor oncology, stable disease is important. We are trying to keep the disease under control, so the entire disease control rate was almost 70%. With metastatic disease, it was slightly less. It was 47%. So, for all practical purposes, about half the patients are responding. But what’s also more important, particularly with immunotherapy, is durability of response. How long can we keep the disease under control?

Shubham Pant, MD:If there’s a quick response and then it kind of grows, there’s durability of response.

Nikhil Khushalani, MD:That is correct. So greater than 50%, I think, if I recall my numbers correctly, about 57% of patients had a response that was durable, at least based on definitions of the trial, which was a response lasting for at least 15 weeks. These data were very recently updated at the European Society for Medical Oncology where, again, in the phase I portion, now the median survival as of January 2018 had reached approximately 12 months. And the durability of response is still standing out and still being seen. So my hope is that this durability will continue with further follow-up on these patients, which I think is critically important.

What it doesn’t answer is, how long should you treat these patients? With melanoma, we are now recognizing that depth of response is important, and for patients who achieve a complete response and have had a defined period of treatment, we may be able to consider stopping therapy in those patients. Whether the same applies for CSCC, I think it is still too early to tell. But in the phase I portion of the trial, the treatment was continued for 48 weeks, with cemiplimab being delivered every 3 weeks. And in the phase II portion of the trial, the option was to go up to 96 weeks, or progression, or the development of toxicity.

Shubham Pant, MD:So 1 to 2 years.

Nikhil Khushalani, MD:Yes, that is correct.

Shubham Pant, MD:So 1 year and then you could go up to 2 years in the phase II?

Nikhil Khushalani, MD:That is correct.

Shubham Pant, MD:So after 2 years, are there any patients who stop therapy, or is it just too early to tell?

Nikhil Khushalani, MD:I think it’s too early to tell. I have not seen that data yet, but I’m sure that will be coming out very soon. Because this is in some ways the first-of-its-kind study. This is the most mature study in CSCC to date. There are other anti—PD-1 antibodies that are being investigated, for example, pembrolizumab in a KEYNOTE study; nivolumab; then an anti–PD-L1, atezolizumab. So I anticipate a lot of these may potentially replicate what we have seen with cemiplimab. I don’t think this is going to be just 1—and that’s my personal opinion—I don’t think it’s going to be only 1 drug that will eventually be effective in this disease. I think we’ll have a cotter of drugs targeting the same pathway. And we will just need to figure out what our personal preferences are in terms of how we are going to use these drugs. In terms of very early data with pembrolizumab, a lot of this was presented at ASCO [American Society of Clinical Oncology] 2018, showing some very similar response rates, all ranging from 40% to 60%. But again, small numbers. Small cohorts of patients.

Transcript edited for clarity.