Cancer cluster causes seem obvious, so why are they so hard to prove?

A diagram illustrating a common thread among different people in varying regions
It may seem like a simple case of cause and effect when people who live in the same area, work at the same job site, or play on the same sports team develop cancer, especially if it’s the same type of cancer. But very few of the more than 1,000 suspected cancer clusters reported to state and federal health officials each year are confirmed to be true cancer clusters.

Every so often, a headline may leave you scratching your head. Like the April 30 Washington Post story about 18 people, three of whom are friends, all of them concentrated in two locations in North Carolina and Alabama, who were diagnosed with a rare form of eye cancer, ocular melanoma. Those involved are convinced their cancers must have a common link, but an investigation failed to find a concrete cause, nor a source tying the cases together. The news left many dumbfounded. “It’s just hard to believe that there’s not a common thread here,” says Pat Cotham, a county commissioner in North Carolina.

It may seem like a simple case of cause and effect when people who live in the same area, work at the same job site, or play on the same sports team develop cancer, especially if it’s the same type of cancer. Some may assume pollution is to blame. Others may suspect toxins in the ground or walls. But proving the connection is often difficult, if not impossible. In fact, very few of the more than 1,000 suspected cancer clusters reported to state and federal health officials each year are confirmed to be cancer clusters, defined by the U.S. Centers for Disease Control and Prevention (CDC) and the National Cancer Institute (NCI) as a “greater-than-expected number of cancer cases that occurs within a group of people in a defined geographic area over a period of time.” Most of the time, the evidence needed to establish a connection doesn’t materialize, either because of outside variables, the limitations of science, eternal mysteries of the world we live in, or because there just isn’t one to find.

Identifying clusters: It's complicated

One example demonstrates the confusion around the enduring questions that arise when seemingly obvious links can’t be confirmed. Last summer, World Series champion Darren Daulton became the fourth Philadelphia Phillies player to die of glioblastoma, an aggressive form of brain cancer. Many people, especially fellow former baseball players, wonder about what seems to be an obvious link between brain cancer and baseball—or, at least, between brain cancer and older versions of synthetic turf used in stadiums like Veterans Stadium, which the Phillies called home from 1971 to 2003. But researchers suspect the cluster is probably a coincidence, saying their investigation did not uncover enough evidence to support the belief that the cancers were caused by a single, shared source. The deaths, they added, fit cancer patterns that occur in the general population: Glioblastoma is more common among men than women, and those at highest risk are between the ages of 45 and 70.

While frustrating, the investigation’s results are not unusual in the world of cancer clusters. In some cases, groups of cancer occurrences are natural because the disease is so common—perhaps more common than many realize. In the United States, nearly one in three people will develop cancer during his or her lifetime. Think of a pool cue breaking up a group of balls in a game of billiards. Some may land in the far corners of the table, while others cluster together. Like cancer afflicting neighbors or co-workers, some experts say, it’s completely random.

And cancer is a complex disease, categorized into more than 100 different types, each with its own risk factors, symptoms and causes. The disease often takes years or even decades to develop. Another factor complicating the quest to prove clusters: People don’t always stay in the same place. 

The reasons cancer clusters are so difficult to prove are numerous, and include the possibility of random chance, duration of exposure, and the possibility that different people are moving in and out of an area or having different amounts of exposure.” - Pamela Crilley, DO, Chair of the Department of Medical Oncology at Cancer Treatment Centers of America® (CTCA)

That’s why many cancer clusters that investigators are able to confirm are found in the workplace, rather than in residential communities—because they are often easier to prove, says Asha Karippot, MD, Medical Oncologist and Hematologist at our hospital in Tulsa. When co-workers who have breathed the same fumes, or drank from the same tap, or handled the same material, over time develop the same type of cancer, the source is confined to a smaller space, and more readily verifiable. The “over time” aspect is another key, both to how cancer forms and confirming a cluster link. Environmentally linked cancers occur most often when people—in the case of cancer clusters, the same group of people—are exposed to high, repeated doses of carcinogens. For example, researchers studying cancer clusters in the 1960s discovered a link between malignant mesothelioma and asbestos, a fibrous mineral then common in manufacturing, industrial and consumer products. Then, in the early 1970s, when workers in the same chemical plant were diagnosed with angiosarcoma of the liver, investigators determined that the rare cancer was caused by the workers’ exposure to vinyl chloride, now known to be a major risk factor for their disease.

Unrelated factors cloud the picture

Other factors, such as smoking, obesity and age, may make cancer clusters even more difficult to prove. “Variables like smoking or obesity can contribute to the appearance of a cancer cluster when one actually doesn’t exist,” Dr. Crilley says. Not everyone who is exposed to a carcinogen will develop cancer—some people are more or less susceptible depending on their genes—while others will develop the same cancer, even if they were not exposed to high-risk toxins. Even when cancer-causing agents are found in a single neighborhood, some residents may be exposed to the substance while others aren’t.

For example, not all the former Phillies players who died of brain cancer played together at Veterans Stadium or stayed with the team for the same length of time. The Philadelphia Eagles also played and practiced at the stadium, but the National Football League hasn’t reported an unusual occurrence of brain cancer among its players. “It’s often difficult to scientifically prove the presence of a cancer cluster,” says Anthony Perre, MD, Chief of the Division of Outpatient Medicine at CTCA®. “It’s very possible that the cancer cluster does exist. It just can’t be proven.”

In order for a true cancer cluster to be confirmed, certain criteria must be met:

  • A greater-than-expected number of cancer cases must occur in a specific population.
  • The cancer cases must involve the same type of cancer or different types of cancer that have the same cause.
  • The cancers must all be primary cancers, meaning the disease originated in the same area of the body, not in organs where it may have spread.

According to a set of defined steps health officials take in investigating a cancer cluster, they gather information, including:

  • The geographic region where the cases occurred
  • The number of cases involved
  • The type of cancer involved
  • Demographic information about the people impacted, such as their age, gender, race and occupation

Sometimes, officials determine further investigation isn’t needed, such as when the cancers:

  • Are among blood-related family members, especially if the cancers are known to have a strong genetic link
  • Are different types that are not caused by the same factors
  • Are the same type but only occur in a few people, especially among those at high risk for that cancer type
  • Occur among people who didn’t live in the same place at the time the exposure would have occurred

If more investigation is needed, officials compare the gathered information with similar census and cancer registry data—such as a group with the same age and gender. The picture is even fuzzier when you factor in the subjective decisions officials make when defining the key elements of a cluster, such as its geographic borders and who it may have impacted. That means that, in some cases, two health departments could arrive at two different conclusions about whether the cancer cluster is real or coincidental.

In 2015, when a suspected cluster of brain cancer was reported in the small town of Belews Creek,  the North Carolina Central Cancer Registry investigated cancer cases in counties with coal ash storage facilities. Most of the town’s population relies on wells for its drinking water, and many blamed a storage pond used by Duke Energy to collect waste from its nearby coal-fired power plant. The investigation, however, couldn’t confirm that the incidence of Belews Creek cancers was any higher than in other, comparable areas of the state. But critics say the study unfairly used county-wide numbers, which may not accurately reflect, for comparison purposes, the number of cancer patients living near the coal ash pond. In December 2017, environmental and civil rights groups filed a lawsuit against Duke, alleging it polluted the community’s water supply. The case is still in court.

Science is limited, but improving

Those who are frustrated by the lack of conclusions on cancer clusters should also consider the limitations of science, experts say. Often, researchers lack the smoking gun needed to establish a connection. For instance, if many believe asbestos is to blame but all traces of the mineral have disappeared over time, researchers have no evidence to draw from. Or, if some suspect their well water is contaminated by a cancer-causing substance, researchers may not be able to pin down exactly who drinks the water and who doesn’t. It doesn’t help that most cluster investigations involve a small number of cases, which may make it harder for researchers to draw statistically significant conclusions. “Many reported cancer clusters just don’t include enough cases to allow investigators to prove that there is a statistical difference from the expected number of cases,” Dr. Perre says.

But experts say technology is helping to clear up at least some of the picture. In the past few years, scientists have developed more sensitive tools to measure exposure, and the CDC and the NCI are now using software designed to analyze geographical cancer data to determine if clusters are statistically significant. National cancer registries, which help establish a baseline of cancer incidence and spot anomalies more easily, are becoming another important resource for researchers in investigating clusters. “There remain many challenges to proving a cancer cluster, but science continues to advance and epidemiologic studies continue to improve,” Dr. Crilley says.