New analysis finds environmental factors, not random mutations, contribute to majority of cancer risk

Researchers publishing in Nature calculated that 70-90% of the risk of developing cancer can be attributed to extrinsic, environmental factors such as ultraviolet radiation and carcinogens rather than intrinsic factors such as random DNA mutations. The researchers evaluated data from a recent study which showed a strong correlation between lifetime cancer risk in certain tissues and the potential for random mutations in the stem cells of those tissues, the so called ‘bad-luck’ cancer hypothesis. Using additional data modeling, the researchers concluded that only 10-30% of cancer risk is contributed by intrinsic factors.


Dr. Andrew D. Maynard, Director, Risk Innovation Lab, Professor, School for the Future of Innovation in Society, Arizona State University (webpage): 

Expertise: Risk assessment, science communication, environmental health policy, and responsible development and use of emerging technologies.

“In this paper, Song Wu and colleagues set out to challenge the “bad luck” hypothesis that suggests some cancers are dominated by uncontrollable random cell mutations, rather than controllable environmental factors.  Using four separate approaches, Wu and colleagues demonstrate that many observed cancer rates can be convincingly understood by assuming environmental factors have a substantial role to play.

Each of the methodologies used by the research team has its assumptions and limitations.  Yet together, they build a picture of a complex interplay between internal random biological processes, and external factors, in the formation and development of cancers.  Their analysis supports conventional thinking around cancer formation and development, and casts doubt on the idea that some cancers are relatively independent of how you live and what you’re exposed to.

“In part, the study is a sensitivity analysis of the “bad luck” hypothesis, testing its robustness under different assumptions. Their results indicate that the “bad luck” hypothesis is rather fragile.

“The authors also suggest an alternative approach to differentiating between the role that random biological events and exposure to external agents play in causing cancer.  By assuming that, for any cancer type, the lowest observed lifetime risk will represent the least influence from external factors, they were able to estimate the relative importance of “bad luck” versus controllable factors outside the body.  They found that, in the majority of cases, environment is influential in determining cancer risk.

“The analysis is not definitive, and there are a number of assumptions made that may influence the conclusions they draw.  However, their conclusion that external factors are influential in determining cancer risk for many cancers is supported by observations, and is strengthened considerably by the methods they use.

“By using multiple methods and combining the results, Wu and colleagues have built a compelling case that many cancers are caused by more than just “bad luck”.”


Dr. Giles Hooker, Associate Professor, Department of Biological Statistics and Computational Biology, Cornell University (webpage):

Expertise: statistical methods, machine learning, dynamic systems. Dr Hooker is an advisor to, a project to improve statistical literacy by the American Statistical Association and Sense About Science USA

“This study adds to the discussion of the “bad luck” hypothesis in cancer: that most cancer is due to ‘intrinsic’ risks, random mutations during cell division, with tissues that undergo more cell divisions more likely to produce cancer in the process. This study proposes an alternative explanation: that environmental risk factors may produce most mutations during cell division, meaning tissues that produce more cell divisions are also more susceptible to external risks.

“These are interesting calculations and give us a sense of the possible range of effects. However, they rely on a very simplified model of cancer mutation and the resulting numbers should at best be regarded as ballpark estimates.

“The authors attempt to separate intrinsic and external risks by producing a baseline, using tissues with lowest lifetime cancer risk relative to their number of cell divisions. Any tissues with cancer risk above this baseline is suggested to be associated with environmental factors.

“By using only the lowest risk cancers and assuming that mutation rates are the same for all tissues, the study maximizes the risk attributed to environmental factors. However, we don’t know how tissues differ in their intrinsic mutation rates. Thus in the authors’ model, a tissue that is more prone to mutations may still be categorized as having high environmental risk. Intrinsic risks could be either under- or over-estimated by this study, depending on how much they vary between tissues; we have no data to evaluate this.

“The study also examines expected lifetime risk of cancer based on estimates of intrinsic mutation rates and shows that they are lower than what they find from published reports. These calculations are based on the simplifying assumption that mutations are statistically independent, whereas if one mutation makes the next more likely, then lifetime risks will be larger than the authors estimate.”


Dr. Jian-Min Yuan, Professor and Arnold Palmer Endowed Chair in Cancer Prevention, Department of Epidemiology, Graduate School of Public Health, Associate Director for Cancer Control and Population Sciences, University of Pittsburgh Cancer Institute (webpage):

Expertise: Environmental exposure and gene-environment interaction in the etiology of cancer

“This Nature report presents a strong counterargument to a report published earlier this year in Science* which began a debate about whether the majority of people who develop cancers do so because of “bad luck” – random mutations arising during the DNA replication in normal stem cells. Using an alternative approach, these authors estimate that up to nine-tenths, rather than one-third as reported earlier, of lifetime cancer risk is due to extrinsic risk factors such as exposures to carcinogens, ultraviolet radiation, and viral infection. The current epidemiological evidence strongly supports an important role of environmental factors in the development of cancer. For example, people who stop smoking at 55 years would cut their lung cancer risk by half compared with those who continue smoking by 85 years of age. HBV vaccine has resulted in the reduction of hepatocellular carcinoma incidence by 70%. These results demonstrate that a large proportion of cancer is caused by environmental factors and are preventable if their underlying environmental causes are identified.”

“The development of cancer is a multi-factorial, multi-step process that we are yet to fully understand. Humans are constantly exposed to a myriad of environmental compounds and carcinogens derived from contaminated food and water, and polluted air, among other sources. However, not every individual exposed to such environmental carcinogens develops cancer, just like not every heavy smoker would develop and die from lung cancer. The individual variability in susceptibility to cancer is determined by other environmental factors such as the exposure to protective factors that can detoxify carcinogens, or by intrinsic factors such as enhanced repair of damaged DNA.”


Declared interests (see GENeS register of interests policy):

Dr. Andrew D. Maynard: AM receives funding through the Michigan State University Center for Research on Ingredient Safety and does not have have any interests directly related to this study.

No further interests declared



Substantial contribution of extrinsic risk factors to cancer development‘ by Song Wu et al., published in Nature on Thursday 16 December 2015

Please feel free to leave your comments below, but be aware that by doing so you agree to our Terms & Conditions.