Avastin and Breast Cancer: The Median is Not the Message

The New York Times ran an op-ed yesterday by oncologist Frederick Tucker urging the Food and Drug Administration to stand by its decision to remove the breast cancer treatment approval for the drug Avastin. That drug has been on the market since 2004 with an FDA approval for the treatment of colon cancer and non-small cell lung cancer, and it has since been approved to treat a handful of other cancers. In 2008, FDA approved it for the treatment of breast cancer after initial testing indicated that it slowed tumor growth, though without the benefit of lengthening the overall survival time. In short, the average breast cancer patient taking Avastin didn’t live any longer than those using other treatments, but many patients reported better quality of life while taking Avastin because of the improvement in so-called “progression-free” survival. So, FDA approved Avastin conditioned on the manufacturer’s agreement to conduct two additional clinical studies further examining the progression-free survival end-point.

Both of the subsequent studies also showed no increase in overall survivability, and only modest gains in progression-free survival. That wasn’t enough for the FDA, however, which decided last July to revoke Avastin’s approval for breast cancer because of its side effects. The product will still be labeled for the treatment of other cancers, and doctors may continue to prescribe it “off-label” for breast cancer. But without an approval, the manufacturer cannot promote Avastin in any way for the treatment of breast cancer, meaning many doctors and patients will be unaware of its availability as a treatment option, and most health plans (private insurers or government-run programs) won’t pay for it. The manufacturer appealed FDA’s decision and will ask the agency to reconsider at a hearing scheduled for next month.

Technocrats, including Dr. Tucker, have applauded the decision, arguing that Avastin doesn’t work very well, and that it has significant side effects. On the other side are patient advocates and thousands of women who have benefited from the drug, who argue that it’s unfair for the agency to take away an option for patients who are at risk of dying. “We want to be sure that women who are using Avastin, and for whom it is working, can continue to have access to it,” said Elizabeth Thompson, president of Susan G. Komen for the Cure. The real question ought to be, who gets to decide?

The biggest weakness in the FDA’s case is its reliance on average or median effectiveness. What averages don’t reveal, however, is the fact that some patients respond to a treatment option much better than others. Particularly when it comes to a life-threatening illness, average or median survivability too often masks the fact that some patients respond very well indeed. Pop scientist Stephen Jay Gould made this point quite eloquently in a 1985 essay published in Discover magazine, titled “The Median Isn’t the Message.” In 1982, Gould was diagnosed with a rare form of lung cancer, which he discovered had a median survival time of just eight months. His doctor had given up on him because of the dire prognosis, but Gould actually lived another 20 years, and then died of a different type of cancer. In his Discover essay, Gould wrote:

[A]ll evolutionary biologists know that variation itself is nature’s only irreducible essence. Variation is the hard reality, not a set of imperfect measures for a central tendency. Means and medians are the abstractions. Therefore, I looked at the mesothelioma statistics quite differently – and not only because I am an optimist who tends to see the doughnut instead of the hole, but primarily because I know that variation itself is the reality. I had to place myself amidst the variation.

When I learned about the eight-month median, my first intellectual reaction was: fine, half the people will live longer; now what are my chances of being in that half. I read for a furious and nervous hour and concluded, with relief …  I knew how to read the data properly and not despair.

Just like Gould’s doctor, the FDA and its technocratic supporters are giving up on cancer patients because of their slavish obsession with median response rates. Sure, on average, Avastin may not extend a patient’s life expectancy, but you don’t have to be Garrison Keillor to realize that some people really are above average. The expected value, a priori, of the treatment for any given patient does indeed appear very small, so you can understand why a health plan would balk at paying the very high price for Avastin (which varies from about $56,000 to $96,000 per year). But making cost-effectiveness calculations isn’t FDA’s job — nor does the Food, Drug and Cosmetics Act even permit the agency to consider the price. Consequently, the FDA had no business removing the breast cancer indication from Avastin’s label, and anyone who values freedom and self-determination should hope the decision is reversed next month.