Bad luck and cancer risk


Those of us at Epicentre have long appreciated that bad luck plays a part in the risk of getting a cancer. This conclusion seems to be unavoidable once it is understood that a cancer results from a series of genetic mutations or other deleterious genomic events that occur over time within a single cell lineage, including previous normal progenitor or stem cells. The chance of such a series of events occurring in one cell lineage is extremely small, although the number of cells in any organ is extremely large.

Many genomic errors that occur arise during cell division, and a report published in Science at the start of 2015 had a surprising finding that the correlation between lifetime risk of cancer and average number of stem cell divisions was extremely high (0.81 to be exact) (1). This seemed to be indicating that cancer risk of any tissue was largely driven by the number of stem cell divisions, i.e. primarily by bad luck, leaving little room for the effects of lifestyle, environment, or genetic susceptibility. Several follow-up reports pointed out various problems in the Science findings (e.g. 2-5) including the use of cancer rates from the US from one point in time. If we take a measure of lifetime risk from another population, especially one that differs dramatically in its cancer burden with the US, the authors would have obtained very different measures of correlation.

As a bookend on the subject at the end of 2015, Nature has published a paper that seeks to take another look at the proportion of cancer risk due to bad luck (6). By using different methods, the authors come to an alternative conclusion that the majority of cancer risk (and over 70% for most cancer sites) is due to external (and therefore potentially avoidable) factors. Our interpretation of the modelling approach used by Wu et al is that it is not conclusive on the subject. For example, a re-analysis by Wu and colleagues of data on lifetime cancer risk and stem cell turnover, or total tissue cell divisions, commits the same fault as the initial Science paper by assuming lifetime cancer risk in the US is representative of cancer risk everywhere else. This we know to be untrue.

In the end, we believe that the strongest data to consider when trying to estimate the role of chance in cancer risk are not stem cell division rates (which are often poorly known) but the extent to which cancer differs across the world in different populations (for which we have very good data). If there is a component of cancer risk due to ‘bad luck’ it should be the same wherever we live. This was the rationale used by Doll and Peto back in 1981 when they compared cancer incidence rates in the US (38 cancers among men and 40 cancer sites among women) with the lowest incidence rates for each cancer found elsewhere (7). This resulted in about 75%-80% of cancer in the US at that time being potentially avoidable, although probably more given that some of these very low cancer incidence rates are unlikely to be just due to bad luck and may include an avoidable component as well. The conclusion that most cancers are avoidable is similar to that of Wu and colleagues, and is an important one, as it means there are a lot of causes of cancer that we still do not understand. Heading into 2016, Epicentre will be picking up on the stories and the studies that help us to fill these missing gaps on cancer causation.



  1. Tomasetti C and Vogelstein B. Science 347, 78-81 (2015)
  2. Wild C et al. Science 347, 728 (2015)
  3. Potter JD and Prentice RL. Science 347, 727 (2015)
  4. Gotay C et al. Science 347, 727 (2015)
  5. Song M and Giovannucci EL. Science 347, 728-9 (2015)
  6. Wu S et al. Nature. Published online 16 December 2015
  7. Doll R and Peto R. JNCI 66, 1191-1308 (1981)
Paul Brennan

Paul Brennan

Paul Brennan is head of the Genetics Section at the International Agency for Research on Cancer, and also one of the EPIC principal investigators. Most of his work focuses on identifying new biomarkers and genes that are involved in some common cancers and how this knowledge can inform cancer prevention, either through identification of new risk factors or early detection.

Research Interests: Cancer aetiology and prevention.
Paul Brennan

Share this article Facebooktwittergoogle_pluslinkedinmail
Follow us on Twitter twitter

Leave a comment