Treatment Inequalities Between Men and Women With Head and Neck Cancer

While head and neck cancer is more common in men than in women, women may actually be less likely to receive aggressive treatment for the disease, which could affect outcomes.

We spoke with Jed A. Katzel, M.D., a medical oncologist at Kaiser Permanente in Santa Clara, CA, about his recent research that found important differences in treatment patterns between men and women.

Can you give some background on your study?

Our group previously investigated head and neck cancer in northern California, looking at demographic and incidence rates by HPV status. On this trial, we wanted to use a validated mathematical model called the GCE model (or generalized competing events model) to evaluate patients more likely to benefit from aggressive therapy. The GCE model also allows us to identify group populations or groups of patients that may potentially be undertreated or over treated in clinical practice.

This GCE model is very exciting. We think it will be very informative. It compares the risk of dying from cancer with the risk of dying from other serious medical problems. It gives a ratio of cancer death over non-cancer death. For example, if the ratio is greater than one, that indicates that the patients' risk of dying from cancer is greater than the risk of dying from other serious medical problems.

On GCE analysis, we found that women were more likely to die from head and neck cancer than other causes compared to men. This was an informative new finding that we look forward to investigating further.

Are there any theories or rationale about why that difference exists?

We will need further investigation to fully answer that question. But we did find on GCE analysis that women were more likely to die from head and neck cancer than from other causes, even when controlling — when the model takes in to account – things like tumor location and stage.

But we do point out that there was a statistically significant difference in the tumor location in our study. In particular, the rate of oropharynx cancer for women in our cohort was 38 percent, while the rate of oropharynx cancer in men was 55 percent. And that's important because we know that oropharynx cancer is more likely to be HPV-related. We know that HPV-related head and neck cancer has a better prognosis and a better response to therapy.

So, we looked forward to investigating this further in a chart by chart review, as well as in the upcoming

, that will actually be using this same model, the GCE model, in a prospective manner.

What can you say about the difference in treatment between men and women?

Another key finding that we looked at in our trial was the difference in terms of intensive therapy. We did notice on a different logistic regression analysis that women received less intensive chemotherapy and radiation therapy compared with men, even when controlling for other factors, like age, smoking history, alcohol history, comorbidities and site of the primary cancer.

We do point out that there were more men in our trial than women. And that's typical in head and neck cancer because we know that head and neck cancer is more common in men. We want to look into this further, though. We plan to review our cohort, the women in our cohort, in a chart-by-chart manner to fully understand why some people may elect to have or decline potentially burdensome and aggressive treatments.

Are there other groups that you're interested in doing this type of analysis on, besides head and neck cancer, to find other areas where there is this type of disparity?

The model that we used, the GCE model, is validated in head and neck cancer. But it's also validated in another number of cancers. It's been validated in prostate cancer, endometrial cancer and breast cancer, along with head and neck cancer. But we see that the value in this kind of analysis may extend well beyond these cancers, and potentially to other disease types as well.

Are there any next steps to move this model forward?

The GCE is a validated mathematical model. It has this potential to identify groups that can be under-treated or over-treated in clinical practice, so it's a very useful tool to help identify differences within patient populations. That was one of the exciting uses that we found in our trial, was to use this type of a model on a large population in northern California.

Is there anything else that you want to mention?

We do want to point out that we look forward to investigating this further, like we said. And we're also very excited about the upcoming NRGH N004 clinical trial because they're actually going to evaluate every patient when they enroll in the clinical trial, and give them prospectively a GCE score, so this will lead to much more information in the future.