There is curious comfort in believing that a cancer diagnosis is just a bad hand. But in the more than 50 years since the Surgeon General’s 1964 report on smoking and lung cancer, we’ve absorbed a steady flood of findings and speculation about what causes cancer. Questions lurk: Did something do this to me? Did I do this to me?
For many, the answer to both those questions may be a fairly definitive “no,” according to a recent study that has stimulated intense discussion in the oncology community.
In January of 2015, Bert Vogelstein and Cristian Tomasetti published a small paper in Science, one of the world’s foremost scientific journals. Vogelstein is a pioneering cancer biologist who discovered mutations in the p53 gene in colon cancer, which led the way to a deeper understanding of cancer as a genetic disease. Tomasetti is a mathematician and a specialist in the mathematical modeling of cancer.
Working together at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Tomasetti and Vogelstein sought to explain why some organs in the body develop cancer more frequently than others. They compared the division rates of normal stem cells from different parts of the body to cancer rates in those organs, and saw a pattern tying greater total number of divisions over a lifetime to cancer. They concluded that some cancers arise simply by chance — more cell divisions beget more opportunities for random mistakes. Factors like carcinogens or inherited gene mutations may also be in the mix, but the bottom line, Tomasetti and Vogelstein concluded, is that randomness is at the heart of most cancer types.
It’s a huge scientific leap, but many cancer patients won’t be surprised. You have two breasts, two lobes of the lung, and yet cancer arises in just one. That’s how most of us think about statistics — a flip of the coin. It won’t earn you a Ph.D., but there is some truth there.
Faced with Tomasetti and Vogelstein’s bad-luck hypothesis, the world of oncology erupted. Science ran a series of rebuttals and objections. Numerous papers populated other journals and disputed their methods and conclusions. Some in the popular press misread and misreported the findings.
The controversy won’t be resolved anytime soon, but raises a fair question: What exactly do we know about cancer and what causes it? A lot more than we did 50 years ago. There are about 200 kinds of cancer, and more subtypes are identified every year. We understand many of the underlying genetic malfunctions. And we have a voluminous catalog of cancerous agents and cancer-promoting scenarios.
Yet, for most individual cancer patients, the most fundamental question remains unknown — and perhaps unknowable. “Epidemiologists do a pretty good job of predicting things for a group,” says Katherine Reeves, a breast cancer epidemiologist at the University of Massachusetts at Amherst. “We don’t do as good a job predicting things for an individual. There is just no way to say: ‘Mrs. Jones, your water bottle caused your cancer.’”
Before the invention of cigarettes, lung cancer was so rare that doctors were surprised to encounter it at all. In 1912, a German physician first suggested that smoking was behind the rise of the disease. Then, in studies published in 1939 and 1943, German physicians connected the first dots. By 1954, epidemiologists felt they had proven the link “beyond a reasonable doubt.” The tobacco companies fought off government action, but in 1964, the Surgeon General began to push back. If anybody knows anything about cancer today, it’s that smoking causes lung cancer.
Still, lung cancer kills an estimated 1.5 million people per year globally, and the consensus is that around 95 percent of those deaths are preventable — partly caused by air pollution and asbestos, but mostly due to cigarettes. But there remains a devilish flip side: Most people who smoke don’t get lung cancer. That’s because cancer is an interplay, a warped dialogue between genes and the environment.
Lung cancer often arises from a series of genetic errors initiated by exposures to carcinogens. Among the known culprits: asbestos, radon gas (from decaying radioactive bedrock) and the combustion products of tobacco, diesel, coal and even cooking oils. The errors accumulate as cells divide to keep our systems operational, and so are passed down as the body ages. Dozens of such errors are typically found encoded within lung cancer tumors.
With or without carcinogens, the continual process of copying of DNA – the entity responsible for directing cell activity — leads to some level of error. The difference is how well your body repairs those errors — in essence, how well your genetic spellchecker works.
“We all inherit DNA repair processes from our ancestors,” explains Lynn Tanoue, medical director of the Yale Cancer Center Thoracic Oncology Program. “If you have really good repair mechanisms, you may be able to handle multiple insults like cosmic radiation, cigarette smoke, diesel fumes, indoor or outdoor pollution — any exposure to any carcinogen,” she says. “If I do not have not good repair, I may not be able to handle any of them.”
Some genetic errors run much deeper, mutations that have been encoded in human DNA for untold generations. Perhaps best known among these are the BRCA mutations that can lead to breast and ovarian cancer. As genetic research advances, more will likely be identified. BRCA is a high-risk mutation, but others are likely to be more subtle — combining with other genes, environmental factors and chance before leading to cancer.
A Simple Twist of Fate
The bad-luck hypothesis has its origins in stem cells, those magical super cells that seem to hold so much medical promise. These are the master copy from which most new cells are made, and we’ve learned that stem cells from different parts of the body divide at different rates. Because each division is another opportunity for genetic error, Tomasetti and Vogelstein wondered what these division rates might tell us about cancer.
Comparing the divisions to the cancer rates for each tissue type, they found a strong correlation — the more division cycles, the higher the cancer rates. Cancer, they say, is an awful lot like a car accident: The longer your trip (the more times your stem cells divide), the more likely you are to have an accident. Road conditions are the environmental factors in cancer. The worse the conditions (the more cigarettes or radiation), the greater the risk of an accident. And the more mechanical problems (your inherited genetic factors, including that genetic spellchecker), the greater the risk of a crash.
And, like a car crash, sometimes an icy road meets a car with balding tires and a sleepy driver. Tomasetti and Vogelstein conclude: “We calculate that two-thirds of the variation (in rates of cancer) is attributable to the random mutations that occur in stem cell divisions throughout a person’s lifetime, while the remaining variation is associated with environmental factors and inherited gene mutations.”
But wait: If 95 percent of lung cancer cases are preventable, then how can this randomness explain two-thirds of lung cancers? It’s an important distinction that even experienced science writers have admitted to getting wrong: Tomasetti and Vogelstein are not saying that two-thirds of cancers are caused by bad luck. They are saying that some tissues get cancer at a higher rate than others, and that two-thirds of this difference appears to be determined by pure chance. It’s an apples-and-oranges thing – some tissues may be much more sensitive to insults, like skin to sunlight.
Nevertheless, the researchers’ conclusions were deeply unpopular with some. An editorial in The Lancet flatly noted the “incompatibility” of their conclusion “with public health evidence and thinking.”
The general umbrage is built upon the fact that we have already uncovered a vast laundry list of factors that contribute to cancer’s development:
> In the United States and other highly industrialized counties, it is thought that a major contributor to between 5 and 10 percent of cancers are viral and bacterial infections, mainly from the hepatitis B and C viruses (liver cancer), human papillomaviruses or HPV (primarily cervical, anal and oropharyngeal cancers), and the bacterium Helicobacter pylori (gastric cancer).
> Another 5 to 10 percent of cancers are strongly associated with known gene mutations that are passed down through families; the proportion is estimated by some to rise to 30 percent when you consider some common gene variants, and some still-unidentified mutations, that can also increase cancer susceptibility.
> Lifestyle factors matter. Smoking, obesity, eating a lot of red and processed meats, not getting enough exercise, getting too much sun and even taking too many vitamins — or not enough — can contribute to the development of cancer. It is also thought that consistently losing sleep can contribute. In how many cancers is lifestyle a major factor? About 42 percent, Cancer Research UK estimated in 2011.
> Demographics matter, too. In 2007, it was estimated that if everyone had the cancer burden of the most educated (a frequent proxy for socioeconomic status) — the number of cancer deaths between the ages of 25 and 64 would drop by more than a third.
> Nineteen percent of cancers, the World Health Organization estimated in 2011, can be linked, with varying degrees of certainty, to exposure to substances capable of sparking genetic changes: workplace exposure to asbestos or coal; chemicals used in manufacturing that enter the air or water supply; fungi or pesticides that contaminate food; substances in makeup, antifreeze, cleansers or pet litter; or (in just 2 to 3 percent of cancer survivors) medical techniques used to treat a previous cancer, such as radiation or certain chemotherapies or immunotherapies.
> Cancer death rates over the last 140 years have been correlated with the rise and fall of galactic radiation.
Tomasetti and Vogelstein aren’t contesting any of this. They are simply saying that chance plays a substantial role too, and if their work stands, they will have uncovered some of the deep biology of cancer.
Their critics have raised valid points and the scientific process is hard at work. But Tomasetti and Vogelstein have certainly struck a nerve: They exposed a mindset within the cancer community that seems almost dangerously proud of its ability to explain all things.
“Our paper brought front and center the probabilistic nature of cancer,” says Tomasetti. “It’s very hard to accept the idea that pure chance, independent of environmental and inherited factors, may play such a large role.”
A Modern-Day Dilemma
Breast cancer is a more complicated narrative than lung cancer, and Jeff Tice, an associate professor of medicine at the University of California San Francisco, looks to Iceland for some clarity. Scrupulous record-keeping by the Icelandic Cancer Registry allowed researchers to analyze increased cancer rates for women who carry the BRCA2 mutation — less than 1 percent of the population. While we know that these women are more likely to develop breast cancer, the study found that, between 1920 and 2000, their risk of breast cancer before age 70 had increased nearly fourfold.
What happened as the century advanced? Richer nutrition led to earlier menstruation, starting the clock on breast development. Improved education pushed back childbearing. “There appears to be a vulnerable period in the breast prior to the first pregnancy,” explains Tice. Women had fewer children, and nursed them for shorter periods. All of those things would have led to more division cycles in the breast, and help explain the higher lifetime risk of breast cancer in more developed countries. “The Iceland example suggests that 60 to 80 percent of breast cancer is probably related to the modern lifestyle,” Tice says. “We probably aren’t going to change that.”
Instead, Tice and his colleagues are building a breast cancer risk model to personalize recommendations for screening. In the model, age is the biggest factor, followed by analysis of a woman’s genetic profile for risk. After that come breast density and family history and then a history of biopsy.
What about weight and alcohol, more modifiable risk factors you may have heard about? Tice says obesity is a pretty consistent risk factor in older women, and may eventually be included in the model. “Maintaining weight or losing weight in women who are overweight appears to decrease the risk of breast cancer,” he says.
Drinking less alcohol is another way to reduce risk. Compared with non-drinkers, women who down two or more drinks per day are about 25 percent more likely to develop breast cancer over the course of a lifetime. And there are other good health reasons to moderate alcohol consumption. Yet, compared to the more powerful factors in Tice’s model, alcohol adds little to the ability to predict who will develop breast cancer.
After treatment, Tice notes that a low-fat diet may decrease the risk of recurrence in the long term, and that’s a healthy option no matter what: Heart disease is still more common in women than breast cancer.
What's in a Water Bottle?
When the phone rang on October 30, 2014, Shana Bernstein saw the unfamiliar number and knew it wasn’t good news. She dumped her young children in front of the television and listened to the preliminary biopsy results: Breast cancer. Small, but soon determined to be aggressive. She recalled her husband from the airport. And then she banged out the first draft on an essay she’d already been planning, riffing on the angst she feels just buying a water bottle.
The ongoing debate about bisphenol A (BPA) and other plastic additives has her consulting her phone whenever she steps into a store. “What role, if any, did the toxins running through our water and food supplies, and our cosmetic and household products, play?” she asks. “I will never know.”
In addition to any toxic risks, she’s a modern woman, late to have children. She’s also Ashkenazi Jewish (another higher-risk group) with a family history of breast cancer. Several years ago, her doctor set up regular surveillance. The diagnosis wasn’t a surprise. “It wasn’t that I had cancer,” she says. “It was that I had it in my 40s” — 42, to be precise.
As for her water bottle, it’s not an easy question to answer. UMass epidemiologist Katherine Reeves is investigating phthalates, a common plastic and cosmetics ingredient. Good lab and animal data suggest that they could cause cancer, but human studies are lacking. “Today, we just can’t definitively say that phthalates cause breast cancer,” she says. Ask me in five years, and hopefully we can give you a better answer.”
Even short of a definitive association with cancer, we know that phthalates and BPA are probably not healthy for us, Reeves adds.
“For people who are either trying to prevent cancer in the first place or who are trying to increase chances of survival, and for living without their disease coming back, focus on healthy eating, healthy body weight, being physically active,” she advises. “There is really good data that all of that can make a difference.”
With one estimate suggesting that 40 percent of cancers are preventable, we can’t afford to ignore that message, so you might want to add these precautions to Reeves’ list of directives: Don’t smoke. Limit exposure to the sun. Eat your fruits and vegetables. Moderate salt and alcohol intake. Read the labels before using household chemicals.
Should you worry about the other 60 percent of cancers, the ones you can’t prevent? Perhaps only from a big-picture perspective. The lesson of smoking is that sometimes the obvious is only part of the answer, and sometimes the biological becomes political. It’s to your benefit that research into public health and cancer biology continues, and that this research informs public discourse.
But there may be little personal profit in reacting with high anxiety when that process suggests there are aspects of cancer we can’t yet control, either as patients or as scientists.
“Our data indicates that, for many cancer types, particularly pediatric cancers, there is nothing that could have been done to prevent those cancers,” says Vogelstein. “It’s uncomfortable for us to not be in control of what the scientists might call stochastics and the public will call bad luck. This is no one’s fault.”