The standard of care for patients who have recurrent metastatic head and neck cancer, if they can’t get a curative local therapy, has been a combination of two chemotherapy drugs and cetuximab and that’s referred to as the EXTREME regimen and it’s recognised to have a lot of side effects. Its median survival has been reported repeatedly to be in the 10-11 month range. We also knew that immune checkpoint inhibitors were active in head and neck cancer and both nivolumab and pembrolizumab were approved in the second line setting. So we hypothesised that these active agents which have potentially lasting effects on the immune response to cancer would have a more profound effect on survival if we used them earlier. We also knew that patients who were PD-L1 expressing, either in tumour or immune cells, were more likely to have an objective response to pembrolizumab. So we hypothesised that using pembrolizumab alone in the biomarker enriched population or pembrolizumab plus chemotherapy in the overall population might be superior to the EXTREME regimen.

So this was a three arm randomised study – pembrolizumab monotherapy for up to 35 cycles; a novel regimen of pembrolizumab plus chemotherapy for six cycles followed by pembrolizumab alone and the chemotherapy being the same as the chemotherapy in the EXTREME regimen; or the extreme regimen for six cycles followed by cetuximab alone.

You mentioned being a biomarker enriched population, how was that being scored?

We used an antibody to PD-L1, the 22C3 antibody, and cells were counted that were either tumour cells, lymphocytes or macrophages that expressed PD-L1. Then a ratio was calculated, the number of PD-L1 positive staining cells to the overall number of cells counted times 100. That was called the CPS and the biomarker populations we looked at were the CPS 20 population and the CPS 1 population.

What kinds of results were seen in these different populations with the different treatments?

Pembrolizumab alone was better than the EXTREME regimen in the biomarker enriched population. If you looked at the CPS 20 population the median overall survival improved from 10.7 to 14.9 months with a hazard ratio of 0.61. If you looked at the CPS 1 population it improved from 10.3 to 12.3 months with a hazard ratio of 0.78. So it appeared as if for PD-L1 expressing cancers, richly expressing or just somewhat expressing, pembrolizumab alone when used first is better than the EXTREME regimen. It did not improve the response rate and it did not improve the progression free survival and we know that there were patients who crossed over from pembrolizumab to chemotherapy when they progressed but in terms of the impact on survival early exposure to pembrolizumab made a profound impact for those patients.
If you looked at the all comers population pembrolizumab plus chemotherapy was also superior to the EXTREME regimen.

How well was it tolerated with these patients?

Pembrolizumab alone was substantially less toxic than the EXTREME regimen if you looked at all grade toxicity, grade 3-5 toxicity or discontinuation for toxicity. Pembrolizumab plus chemotherapy was equivalent to the EXTREME regimen in terms of toxicity of any grade, grade 3-5 toxicity, and there were no new safety signals seen apart from what we already knew about pembrolizumab and what we already knew about chemotherapy.

You mentioned that the response rate wasn’t improved but the duration was.

Absolutely. So for pembrolizumab alone, for patients who had a response to pembrolizumab alone, the median duration of response was over 20 months. That’s profoundly different from what we see with the EXTREME regimen.
So for doctors working in the clinic now, patients entering the clinic with a PD-1 positive squamous cell carcinoma of the head and neck would you recommend this as a first line treatment?
I would. There are a small number of patients who may not be ideal candidates for immune checkpoint inhibitors. We excluded people who required steroid therapy or who had active autoimmune diseases and there are patients who are highly symptomatic from recurrent head and neck cancer for whom the higher response rate that we saw with pembrolizumab plus chemotherapy might be more beneficial. There are going to be forthcoming analyses looking at some of the clinical predictors of response and benefit as well so there’s still more to be learned from this trial.

Speaking of, are there any planned follow-ups to establish impact on overall survival in the long term, five year follow-up, ten year follow-up?

This was an overall survival endpoint. There are still between 10 and 16 patients ongoing on the various arms. Patients who have durable response and are in survival follow-up we will continue to follow.