Expert Calls for a Change in the Staging of HNSCC

Loren K. Mell, MD

The way that cancer, especially squamous cell carcinoma of the head and neck, is staged and risk stratified is inherently flawed, according to Loren K. Mell, MD.

During his lecture at the 2017 OncLive^® State of the Science Summit^TM on Advanced Head and Neck Squamous Cell Carcinoma and Thyroid Cancer, Mell discussed the need to properly assess patients’ competing risks when considering their treatment options.

“We need to be patient with the research and really pin down good evidence,” said Mell. “We are making every effort to really individualize care within the context of evidence-based medicine. These are inherently in conflict with one another because for us to gather enough evidence, we need to pool a lot of people and treat them somewhat similarly so we can make some statement about the data.”

OncLive: What are the biggest issues right now with staging for head and neck cancer?

In an interview that took place during the meeting, Mell, who is an associate professor at the University of California San Diego School of Medicine, explained his research for risk-stratifying patients with head and neck cancer.Mell: What we are interested in knowing is, essentially, which patients are most likely to benefit from more versus less intensive therapy. There are 2 components to that. First—is this therapy more effective? Usually, the gold standard for testing that is a randomized trial. The other component is the specific individual in front of us and their characteristics.

What we are trying to do is come up with a different method and approach that makes more mathematical sense than historical standards. The idea that most people will be familiar with is staging—[patients] often want to know the stage of their cancer. Staging has an interesting history to it. Essentially, it is used for a couple different purposes. One is to try to get a sense of how extensive or advanced the cancer in front of us is, and whether we should do more limited treatment. Or, is it more advanced, where surgery may be off the table? Those people typically range from 1 to 4, with 4 being quite advanced. We know that there are many other factors that we need to know about. Just knowing someone's stage isn't sufficient enough to formulate a treatment plan for them.

A lot of our attention has rightfully been focused on characteristics of the cancer itself. What tumor size is it? Has it spread to the nodes? Has it spread to distant parts of the body? But there are genetic factors that might determine [a cancer’s] behavior. Particularly, if we have drugs that act in one way for a certain mutation, that would be ideal to know.

What led to the need for an assessment like this? Are patients being overtreated or undertreated?

A sort of less appreciated, but equally important, component to it are other factors that are maybe less biologically dramatic but no less important. Age is a big factor, as well as other health problems. These are things that we need to incorporate into our treatment plan. The answer is both. The reason for that is the dominant paradigm that we use to risk-stratify patients is okay but fundamentally flawed. The fundamental flaw has to do with the fact that most of these model patients’ risks. If you try to sort out 100 patients, you say, “These top 25 patients are highest risk and these bottom 25 are lowest risk.”

Those models tend to focus on 1 endpoint, or class of endpoints, called event-free survival. The typical one would be overall survival. That is very simple to quantify and it is unambiguous. It is a desirable endpoint to know, and by and large it works, but where it breaks down is where people have an appreciable risk of dying of something else.

Those models fail because what they do is treat a death from cancer equally to a death not from cancer; those are obviously not the same thing. Someone who has a very poor survival tends to be labeled as high risk. However, imagine a patient where 75% of that risk is attributable to one of their other health problems. That makes no sense because, per the model, you would put them in the same risk pool as someone who has a horribly advanced cancer and needs more treatment.

How would improving this risk-assessment program impact the treatment of patients with head and neck cancer?

If you do not separate those events, you have no hope of solving anything. This is part of the problem with risk-stratification schemes that focus solely on the tumor because, at best, they can only explain part of the problem. Our treatments for other diseases have gotten better. People are aging better overall and our population is aging. We are seeing a lot more older patients presenting with cancer, have many years left to live, and look like they are 50. Yet, our current treatment paradigm tells us that those patients don't benefit from chemotherapy.

Rightfully, they walk into an appointment and ask, "Why would you treat me any differently than someone 20 years younger?" This has become a debate. We would like to apply things where there is evidence and, right now, there is strong evidence that patients older than 70 do not benefit from chemotherapy. That is a hard thing to say, because how do I know that a particular individual won't benefit [from chemotherapy]?

The answer has partly to do with what our research is focused on, which is a different paradigm for understanding how to risk-stratify patients. To do it, you need to break away from the question of, "What is this patient's survival?" to, “What is their relative risk of dying of cancer as opposed to other things?” That gives you very different inferences on how intensely to treat patients.

The converse works, too. [There are] patients who we would normally want to give chemotherapy, and we know right away if their organ function doesn't permit it. However, what if they have a constellation of factors? It just predicts that they are not likely to benefit if you consider their age and their poor performance. There are other factors, too, like a patient's economic wellbeing, their marital status, and their psychiatric wellbeing—these have physical manifestations in a person's prognosis. If we could somehow take these factors into account, we might be able to weigh benefits and toxicities better for each specific patient.

What would be your take-home message for the audience?

In head and neck cancer, it gets quite complicated rather quickly because patients have cancer, but sometimes they are smokers, so their lung function isn't very good. Or, maybe they are older with a relatively slow-growing cancer. The research is trying to distill down a better quantitative approach to managing our patients. Now, we must do the hard work in reverse by trying to fit an individual into a paradigm. It is quite a tricky business, but one that we steadily learn and can sort out.

The other thing that changes the landscape is the whole treatment paradigm. For a long time, we have had the "classic 3"—surgery, radiation, and chemotherapy—but now you have this new kid on the block in immunotherapy. It is a different paradigm for cancer treatment and it has some very appealing options, especially when we talk about older patients whose immune functions are in decline. This is a [topic of] constant revisiting; that’s why we call it research.

What we would like to do is expand this idea beyond head and neck cancer. The way that we stage cancer across the board is fundamentally flawed. However, this is a hypothesis that requires us to test rigorously. Therefore, regardless of what we share, it does take careful study and will, hopefully, stimulate a little more interest in this field, which has long been sidelined in oncology.