Wednesday, November 26, 2008

The Case of the Vanishing Breast Cancer

The most startling medical finding of the week comes from breast cancer researchers. Their study “…raises the possibility that the natural course of some screen-detected invasive breast cancers is to spontaneously regress.” (emphasis mine)

A population study

Per-Henrik Zahl, M.D., Ph.D., of the Norwegian Institute of Public Health, Oslo, and colleagues retrospectively compared the number of breast cancers found in more than 100,000 Norwegian women who received mammograms every two years with a roughly equal number who received only one mammogram after six years. To their surprise they found an increased risk of invasive breast cancer in every group of women over the age of 39 who received mammograms. These women were 22 percent more likely to later develop breast cancer than the control group.

What does it mean?

How do we interpret this bit of information? Does screening identify cancers that were destined to spontaneously regress? Maybe, and unfortunately that would imply that millions of women have received uneccesary treatment.

I'm allergic to being alarmist, but
Isn't it possible that the additional (albeit low level) radiation of the mammograms served as a tumor catalyst?
The late John Gofman, MD, PhD, raised red flags and eyebrows on the subject for decades. A 2005 blue-ribbon panel of the National Academy of Sciences examined hundreds of articles and concluded that no safe threshold exists. I wish I had seen a discussion of the health effects of ionizing radiation, e.g., mammograms in the reports following this publication.

How useful is this new study?
Dr. Jeanne Mandelblatt from the Georgetown University Comprehensive Cancer Center said the study’s design can’t prove the cancer regression hypothesis because it wasn’t a randomized controlled trial. But a randomized trial is highly unlikely to occur: few if any women in the US could be persuaded to forgo treatment once an invasive tumor is identified. Any study so designed would be considered unethical.

Wait, it gets better

In May the Cancer Registry of Norway published a large study which found that five percent of breast cancer tumors appear to double in size in just over a month. The rapidly-growing tumors were found mostly among the younger women in the cohort. Records of nearly 400,000 women aged 50 to 69 were compiled for this retrospective analysis. The study, published in the journal Breast Cancer Research, found detection rates that varied from just 26 percent for a 5 mm tumor to 91 percent for a 10 mm tumor.

The Norwegians are presenting us with a potential dilemna. Rapidly-growing tumors are a formidable enemy, worth extra effort if we can differentiate them from so-called "indolent" cancers. Should younger women be getting screened more often? Current conventional wisdom says no, the additional risks (false positives, additional, cumulative ionizing radiation) do not justify any known benefit.

Data mining for women’s health

Crunching numbers, no matter how well, is not a substitute for truly understanding the natural history of breast cancer. Screening, as we see more clearly than ever, is not prevention. Risk models and mammograms are useful tools mostly because treatment of early stage cancer is so much more effective and easier on the patient than treatment for advanced cancer.

Proceed with caution

We partner with our doctors to make our best-informed decisions, including new data as appropriate. But none of these findings is cause for radically changing our current approach. Individual prevention, screening and treatment decisions are just that – individual.

No comments: