Despite the study in one of the world’s most prestigious scientific journals, and the high-profile media coverage of it: no, cancer is absolutely NOT mostly random and a product of “bad luck.” It wasn’t true when these same investigators published very similar work generating very similar media hype and nonsense two years ago, and it isn’t true now.
That’s what I want to talk about, but must hasten to append the obvious proviso: cancer can, rarely, be or at least seem utterly random. But to call cancer random because a child rarely gets retinoblastoma or glioblastoma, or because a non-smoker gets lung cancer would be like calling injury from violence “random” because of what stray bullets can do to innocent bystanders. In both cases, horrible things have happened for no apparent reason.
Of course, even the most apparently random of terrible fate may not be truly, completely random. We have abused our planet in innumerable ways and riddled our environment with a vast inventory of toxins. Perhaps some of the most seemingly random cancers to affect the human body are related to such transgressions by the body politic. Similarly, a stray bullet and its victim can be utterly random- but such a tragedy is more likely, and more common, in a society with a highly permissive attitude about guns and the profits attached to their exchange.
In such ways, then, cancer can be “random”- but overwhelmingly, it is not.
The evidence we have that cancer is often, even routinely preventable by choices we can make singly (i.e., lifestyle), or collectively (i.e., policy, and the social/environmental determinants of health), is nothing short of incontrovertible. We have clear data linking a number of modifiable exposures, tobacco salient among them, to cancer incidence. We observe enormous variations in cancer rates, both in general and of specific types, among different populations around the world. We even observe an enormous change in cancer rates when the same population, with the same genes, changes its environment and lifestyle.
Now, as two years ago, the researchers at Johns Hopkins examined the correlation between a type of mutation, called replication error, and the incidence of cancer by tissue type. The principal difference between this study and the prior was that they looked only at the U.S. population last time, and this time, the sample was global. With all due respect to these investigators, the other major difference is that this time, they overstated the implications of their research rather more egregiously. Last time, in their study title of January, 2015, these authors only said that cancer was often attributable to random mutations. This time, they go on to assert in their title the implications for cancer prevention.
Specifically, the researchers suggest that since so much cancer is associated with random mutations, primary prevention- forestalling cancer from developing at all- simply isn’t possible. Their advice is, in effect, more cancer screening: colonoscopy, mammography, and such.
I fully support cancer screening, of course, but the study authors’ assertion is nonetheless wrong. For one thing, we simply lack reliable screening approaches for many, even most cancers. For another, the conflation of mutations in tissues for cancers in people is as misleading as it was two years ago.
This entire line of research basically indicates that when cells divide and DNA is replicated, copy errors occur. We knew that. It also indicates that the frequency of those errors correlates with the frequency of cell division. We knew that as well, and it’s why tumors of the brain and heart are very rare, while tumors of the skin, gastrointestinal tract, and reproductive organs are common. The latter are tissues with high rates of cell division, and into the bargain, large surfaces exposed to potential toxins from without or within.
So, the key finding about copy errors in DNA could be translated to a corollary about copies of office memos: the more copies somebody in the office makes, the more likely that an error will turn up in their batch of copies. More copies, more errors.
Why, then, doesn’t this mean what we are now being told it means- that we are all rather helpless victims of replicative happenstance, prone to more frequent cancer in our more frequently dividing tissues? For two reasons at least.
First, there can be a high, or even nearly perfect correlation between mutations and cancer in a given tissue, and yet be massive variation in the absolute frequency of both in tandem with environmental factors and lifestyle practices. So, for instance, there could be a “random” mutation associated with breast cancer 75% of the time in both the U.S., and Japan- but the rate of breast cancer in Japanese women could be only half that of American women. Historically, in fact, that has been the case- although sadly, the rate of breast cancer is rising in Japan with globalization and the subordination of traditional, cultural practices to bacon-cheese burgers, fries, and Coca-Cola for all.
Second, a mutation is not a clinical cancer; not by a long shot. We are all subject to mutations all the time. The tissues that divide most often, and produce the most mutations, are also apt to renew themselves routinely. We slough, and replace, our inner and outer linings- skin, and intestinal epithelium- at a very high rate. Many of the mutations that occur there may be jettisoned accordingly.
When mutations occur in cells not carted out with the trash, the body may nonetheless dispatch them effectively. Among the many crucial functions of the immune system is “immunosurveillance,” prowling the body for rogue cells. The very lifestyle and environmental factors associated with low rates of cancer foster the efficiency of immunosurveillance; converse exposures conspire against it, offering mutated cells an exploitable opportunity.
Lastly, cancer develops through a sequence of well-delineated steps, of which a relevant mutation is only the first: initiation. The second, progression, is again highly dependent on lifestyle practices and environmental conditions. Mutations can occur rather instantaneously, while progression is a process often involving years, even decades. Thus, the greatest cancer prevention opportunities reside at this step, to which the new study is inattentive. A mutation may be necessary for cancer to develop, but it is with rare exception quite far from sufficient.
We might say, I suppose, that ships sink randomly because ships tend to sink in storms at sea, and storms at sea are not under our control. The analogy with cancer is fairly robust. Storms at sea are random to a degree, but we also know where, and in what season, they are most common. Similarly, we know about social, environmental, and lifestyle factors that favor mutation, and impair immunosurveillance. More importantly, although any ship is more likely to sink in a storm than in fine weather, the likelihood of shipwreck varies, of course, with the competency of captain and crew, and the seaworthiness of the vessel.
Mutations are the storms; the human body is the vessel; and we are captains, all. Cancer is, to be sure, a constant threat- like storms at sea. But it is generally no more random than the risk of shipwreck.
For scientists to be heard above the constant cyberspatial din of modern life, there is some insidious pressure to push the implications of research findings as far as the proverbial envelope will bear. For the media to garner attention above the same background static and in ever tighter news cycles, there is rather overt incentive for hyperbolic and even willfully absurd headlines. Since this appears to be rule rather than exception where science and media collide, there is nothing random about it.
In this case, since much the same research generated much the same misrepresentations two years ago- we may reasonably consider the entire episode a replication error.
Director, Yale University Prevention Research Center; Griffin Hospital
Immediate Past-President, American College of Lifestyle Medicine
Senior Medical Advisor, Verywell.com
Founder, The True Health Initiative