Science Matters

(11/2022) Your neighbor tells you her husband has been diagnosed with cancer. That’s bad, but she has worse news. Two other families on your street have received cancer diagnoses. Over the next few days, the neighborhood grapevine reveals still more cancer cases in your area. Scariest of all, you hear that two children in your local school have cancer. And you all live in the same neighborhood.

Welcome to the phenomenon of the cancer cluster.

If you’re like most people, you begin to be afraid. It feels as if a common cause has struck. You wonder whether you or your family have been exposed to it. Surely, you imagine, something is going on.

Could it be that industrial site? Was there a chemical spill? What’s in the water? There’s talk of calling a community meeting and of alerting health officials.

As frightening as situations like this can be — and some Frederick residents have experienced this, most notably near Fort Detrick — apparent cancer clusters are, in fact, very common. And in 99.8 percent of cases, they are nothing to worry about.

Each year across the country, more than a thousand suspected cancer clusters are reported to state health departments. But, according to the Centers for Disease Control and Prevention (CDC), out of every 1,000 suspected cases, only two turn out to be real. That is, only two-tenths of 1 percent show good evidence that there is a common cause. The rest are pure coincidence or can be explained by other factors.

One of the most famous supposed cancer-clusters came to light in the 1980s when it was claimed that women on New York’s Long Island had an abnormally high rate of breast cancer. And, indeed, the rate of cases was 18 percent higher than the national average. The numbers frightened women and prompted formation of advocacy groups that waged loud publicity campaigns. Eventually Congress appropriated $30 million to study the situation.

The conclusion was that there was nothing surprising because there were factors other than location that explained the "cluster."

For example, women on Long Island were more affluent than the national average and as a result were more likely to have their first children at later ages. It is well established that the later a woman becomes pregnant, the higher her chance of breast cancer. Nuns, most of whom never have children, have the highest breast cancer rates of all. Also, Long Island’s alcohol consumption was higher, and that also correlates with breast cancer. So does using birth control pills, which was more common among affluent Long Islanders.

Once these other factors were figured in, the breast cancer rate turned out to be the same as was found in other parts of the country with similar demographics.

Situations like these explain away some apparent cancer clusters. Many more can be chalked up to pure randomness. There is nothing that ensures that the incidence of any disease will be distributed perfectly evenly across the landscape, say one cancer case in every tenth house along every street in town. Instead, as cases crop up, there may be two next door but the next few blocks may have none. By happenstance, an observer may live in a block with no cases while another lives in a block with two or three.

To appreciate this phenomenon, imagine a billiard player breaking the neat triangle of balls at the start of a game. The 15 balls scatter, bounce and come to rest. If three balls stop close to one another, nobody imagines there was something about that spot on the green felt that pulled them there.

Public health experts use several criteria to decide whether a suspected cancer cluster has a common cause. First of all, different kinds of cancer have different causes. So, for example, you can’t lump breast cancer cases with colon or lung. Another factor is whether the number of cases of a given kind of cancer is greater in one defined area than it is in a comparable area elsewhere with the same demographics such as age, sex and, say, smoking history. That’s what explained away the suspected Long Island breast cancer cluster.

Every state has a cancer registry that keeps statistics on what cancers are appearing where and in what sorts of people. From these numbers, specialists can see how many cases of a given kind of cancer are routine in an area. The question for public health experts is whether the number of cases of a given kind of cancer in a suspected cluster is greater than would be expected from the overall average.

Accordingly, a small cluster of a rare type of cancer could be cause for concern while a larger number of common cancers would not be.

Another factor to keep in mind is that all cancers are the result of an accumulation of specific genetic damages or mutations. Several different specific genes, each involved in cell division, must be altered to produce cancer. Tobacco smoke, sunlight, some chemicals and certain viruses can cause such mutations. But so can simple mistakes made when cells divide in the body.

Cell division happens at different rates in different tissues — frequently in the colon, for example, but rarely in the brain. The copying of DNA in the process of cell division is very accurate, but not always perfect, and the more often cells divide, the more likely a mistake will happen in a gene that regulates further cell division. All these factors explain why older people are more prone to cancer — their cells have divided more times than those of a younger person. And they have had more time to accumulate mutations caused by outside agents.

If your neighbor’s husband — the initial cause of that hypothetical concern about a cancer cluster — moved next door just a few years ago, the cause of his disease may have happened long before. Most of his mutations would be the result of things that happened when he lived somewhere else. And they may have been caused by random errors in DNA copying rather than any "chemical."

Check out the other cases in your supposed cluster (which cancer epidemiologists do in suspicious cases), and you are 99.8 percent likely to find that they share few or no common characteristics. You have a cluster without a common cause. Sympathize with your neighbor, but don’t worry about you and your family.

Boyce Rensberger retired to New Midway after more than 40 years as a science writer and editor, mostly for The New York Times and The Washington Post. Write him at boycerensberger@gmail.com.