hen a high-profile study this year concluded that many cases of cancer are the result of “bad luck,” the backlash was as furious as if the scientists had advised the public to enjoy a nice cigarette while lying on a tanning bed. “Problematic” and “dangerously misleading” were among the more polite reactions, especially when news stories reported that two-thirds of cancer cases are due to unlucky, intrinsic biology rather than, say, smoking, sunlight, and unhealthy diets.
Now the battle over whether cancer is largely the result of bad luck (meaning random errors that cells make when they divide) has escalated. In a study published in Nature on Wednesday, a team of statisticians and cancer biologists describes multiple lines of evidence — mathematical, epidemiological, and molecular — that undermine the idea that most cancers are the result of random biological mistakes.
“A majority of cancers are due to extrinsic factors,” said Dr. Yusuf Hannun, director of the cancer center at Stony Brook University and senior author of the new paper.
Built-in cancer risk
Ironically, given the brickbats that flew when the “bad luck” study was published in January, that paper’s authors don’t really disagree.
“We definitely didn’t say that two-thirds of cancers are due to intrinsic factors,” said statistician Christian Tomasetti of Johns Hopkins University, a coauthor of that controversial study. “It’s very clear that environmental factors affect cancer incidence.”
The earlier study, by Tomasetti and eminent cancer biologist Dr. Bert Vogelstein of Hopkins, was, on its surface, fairly simple. It examined whether there is a relationship between how often certain cells divide in different kinds of human tissue and the lifetime chance that cancer will develop in that tissue.
They found a relationship, and a strong one. The more often a tissue’s cells divide, the more likely that tissue is to develop cancer. Brain cells rarely divide, for instance, and brain cancers are rare; colorectal cells divide like crazy and colorectal cancer is common. There are important exceptions, though. Lung and prostate cells rarely divide, but those organs account for a large fraction of cancers; cells of the small intestine divide all the time, but cancer there is rare. Still, the general relationship made biological sense: when cells divide they duplicate their DNA; duplicating DNA can introduce mistakes; DNA mistakes can cause cancer.
Specifically, the Hopkins scientists concluded, two-thirds of the difference in lifetime cancer incidence in different tissues — high in some, low in others — comes from differences in the rate of stem-cell division in those tissues.
That finding was “interesting and possibly important biology,” said Dr. John Potter of Fred Hutchinson Cancer Research Center. “But it was primitive and uninformed epidemiology.”
That was true for two reasons. First, many news reports garbled the study’s findings by saying that two-thirds of cancer cases reflect this intrinsic stem-cell behavior; the study didn’t say that. Also, the last sentence in the paper, published in Science, said, “primary prevention measures are not likely to be effective” for 22 cancers, including melanoma (often caused by ultraviolet light) and esophageal cancer (smoking, alcohol, acid reflux).
That infuriated experts in cancer prevention. But, Tomasetti told STAT, it was a typo. What he and Vogelstein meant to say was that prevention is “not likely to be AS effective” in cancers whose underlying tissue undergoes many cell divisions. That’s a less controversial conclusion. But the dropped “as” was not corrected online for weeks, and readers of the paper edition of Science might never have seen a correction at all.
When the Stony Brook researchers dug into the controversial paper, they found much more to flag than typos, however.
For one thing, the Hopkins paper did not include common cancers such as prostate, breast, stomach, and cervix, of which the last three have major environmental causes. Omitting them can skew the conclusion about the importance of cell division in explaining which tissues are more or less likely to develop cancers.
For another, the new paper argues, cancers from cell-division errors are just the base to which cancers due to other causes are added. Most cancers are more prevalent than cell division can explain, Hannun said, as shown by huge geographical variation in rates of different cancers: breast cancer is five times more common in Western Europe than East Asia, for instance, and prostate cancer is 25 times more common in Australia than south-central Asia. That suggests that cancers are triggered by much more than intrinsic factors like cell division.
Molecular signatures of many cancers also suggest they come from extrinsic hits, such as exposure to carcinogens, and not intrinsic cell processes. Moreover, what the Hopkins scientists called “intrinsic” might not be. “External factors can influence the rate of stem-cell division and the number of mutations,” Hannun said.
From these and other lines of evidence, he and his team concluded “unavoidable intrinsic risk factors [such as cell division] contribute only modestly, less than 10-30 percent, to the development of many common cancers.”
While the two studies seem to be at odds, they’re also talking past each other. The Hopkins study was about relative risks — that is, the risk of cancer in this tissue vs. that tissue. The Stony Brook study is about absolute risks, and the contributions of intrinsic and extrinsic factors.
Cancer experts not involved in either study generally praised the new paper. Several described its use of four different lines of evidence as “compelling” and called it a useful corrective to the meme that most cancer is the result of bad luck rather than carcinogens that can be minimized through either personal behavior or public policy.
“In general terms, the conclusion [of the new paper] is robust,” cancer epidemiologist Paul Pharaoh of the University of Cambridge told Britain’s Science Media Centre. Biostatistician Kevin McConway of England’s Open University said he was “impressed,” calling it “pretty convincing evidence that external factors play a major role in many cancers.”