Transcript:Robert L. Ferris, MD, PhD, FACS: The underlying biology of head and neck cancers is interesting because there are two main subsets. Traditionally, head and neck cancers were induced by carcinogens, environmental exposures, such as heavy tobacco exposure and also heavy alcohol use, and these induced changes in the DNA, genomic alterations that led to cancer development. Over the past few decades, we’ve recognized that another major subset across the world is the virally-induced cancers. In North American and western countries, we have human papillomavirus. In the east, Asians have a high frequency of Epstein-Barr Virus (EBV)-induced nasopharyngeal carcinomas. It’s important to distinguish head and neck cancers as carcinogen-exposed or virally-induced.

Head and neck cancer is quite a heterogeneous disease. We’ve begun to realize that even more recently, the genetic drivers were actually initially characterized by identifying alterations in the tumor suppressor, p53. P53 inactivation, is the most common genetic event in all of human cancers, and head and neck is no different; 60% to 70%, and perhaps more the harder you look, of p53 is altered in head and neck cancers. The rest where p53 is not altered are generally the virally-induced cancers, and they inactivate p53 through non-genetic means, through the expression of viral proteins.

Besides p53, which is the dominant genetic alteration, there are some other scattered genetic mutations in some tumor suppressors, particularly in the Notch pathway. That was characterized through a paper out of the University of Pittsburgh and others at MD Anderson. In HPV-positive head and neck cancers there’s a relatively high frequency, perhaps half with activating mutations in the PI3-kinase gene and, in fact, there are approximately double the PI3-kinase activating mutations in HPV-positive tumors, about 50% as opposed to HPV-negative ones, which have about 25% activating PI3-kinase alterations.

Now, if one looks within an individual’s cancer, you’ll find that not every tumor cell has these genetic alterations. And so that heterogeneity is probably where we see variability in treatment response because the PI3-kinase alterations may only be present in a subset of the cancer cells in the tumor microenvironment, and that’s what we mean by heterogeneity, that there’s variability in essentially when we do these genomic analyses, it’s an averaging across a whole tumor.

We’re beginning to understand that there are subpopulations that we term heterogeneity within a tumor and some of them have different genetic alterations. The heterogeneity is not just genetic though. There are differences in the host immune inflammatory response to the cancer. And, in a certain region of the cancer, you will see infiltrating lymphocytes and immune cells trying to recognize, target, and reject the cancer cells. In other areas of the cancer, you see no immune infiltrate, no inflammatory cells. It’s really an inert portion of the tumor. So that sort of heterogeneity plays a major role in treatment response in the development of cancers and prognosis.

Ezra Cohen, MD: The epidermal growth factor receptor we’ve known about for a long time and its association with head and neck cancer, we’ve known that higher expression of EGFR is associated with a worse prognosis, and that’s a worse prognosis almost irrespective of how the patient is treated, whether it’s surgery, radiation therapy, or chemotherapy. We’ve tried to exploit that mutation with agents that target the epidermal growth factor receptor and what we’ve learned along the way is that there are tumors that genomically upregulate this protein, and they do that in one of two ways; either through true EGFR amplification, so more than one copy of the same gene on the chromosome or an increase in gene copy number, that is replication of that part of that arm of the chromosome or the entire chromosome. Both ways appear to be quite effective at increasing the protein expression. What’s interesting is that when we began to look at more and more tumors, and especially when we began to look at the differences between HPV-positive and HPV-negative tumors, we noticed that this EGFR amplification was exclusive to HPV-negative tumors.

The TCGA analysis was really a broad attempt at molecularly profiling several cancers and eventually squamous cell carcinoma of the head and neck, 500 samples were analyzed by profiling that included sequencing mRNA expression, epigenetics, and proteomics. Now we have a comprehensive catalog of what squamous cell carcinoma of the head and neck looks like at the molecular level. Let me first talk about some of the limitations in TCGA and then we can talk about the relevance to day to day practice. One of the limitations is that because of the nature of the way the specimens were collected, there are relatively few HPV-positive cancers. You have to remember that in order to submit samples to TCGA, one had to submit a fair amount of tumor and normal tissue with some other specimens and accompanying data. That really left the bulk of the submitted cancers being ones that were operated on. For the most part, the primary treatment for HPV-positive cancer isn’t surgery, so most of the samples in TCGA are HPV-negative. There are some HPV-positives and it makes up about 20% of the cohort, and so it’s a substantial number. We have to keep that in mind when we begin to look at the data.

With all that in mind, what we began to see were some really interesting findings. First of all, it became quite clear that HPV-positive and HPV-negative head and neck cancers, at a molecular level, are distinct diseases. They share some common alterations, but for the most part, the alterations are different. HPV-negative cancers seem to alter tumor suppressor genes, most commonly p53, p16, and FAT1, while HPV-positive cancers, interestingly, the most commonly altered gene is PIK3CA and the PI3-kinase pathway.

What does that mean for therapy? It could mean a few things and we’re beginning to explore these right now. We identified what we call actionable alterations, so they were alterations in HPV-positive cancers in PI3-kinase. In fact, almost half of HPV-positive cancers will have a genomic level alteration in the PI3-kinase pathway, and we think that would be fertile ground for future therapies and that’s being explored. On the HPV-negative side, we found a few targetable alterations. EGFR amplification is one, Cyclin D1 amplification occurs in almost a third of HPV-negative cancers, and then there are low percentages of FGF receptor mutations or alterations. Incidentally, FGF receptor alterations also occur in HPV-positive cancers, just different ones. But, we have FGF receptor inhibitors at our disposal, some of them, in fact, commercially available, and those things are being studied in clinical trials right now.

Transcript Edited for Clarity