Is cancer mainly bad luck? NIEHS scientists respond
By Kelly Lenox
A study published January 2 in the journal Science attracted attention in the scientific and popular press alike for its conclusions, which were often summarized as “cancer is just bad luck.” The study, “ Cancer etiology. Variation in cancer risk among tissues can be explained by the number of stem cell divisions,” was conducted by Cristian Tomasetti, Ph.D., and Bert Vogelstein, M.D., both with the Johns Hopkins Kimmel Cancer Center.
Two NIEHS scientists, Clarice Weinberg, Ph.D. and Dmitri Zaykin, Ph.D., published a commentary in the Journal of the National Cancer Institute, providing another perspective on the original study (see summary). “While bad luck may play an important role in carcinogenesis,” they wrote, “the data do not compel that conclusion.”
Taking a closer look
Tomasetti and Vogelstein compared average lifetime incidence rates of 31 site-specific types of cancer with the number of stem cell divisions for the associated tissues, and after plotting logarithms, they found an association between stem cell division and cancer rates. “These results suggest that only a third of the variation in cancer risk among tissues is attributable to environmental factors or inherited predispositions,” they wrote. “The majority is due to ‘bad luck,’ that is, random mutations in normal, noncancerous stem cells.”
Weinberg, head of the NIEHS Biostatistics and Computational Biology Branch, and Zaykin, a lead researcher in that branch, took a closer look at the data and raised three statistical concerns.
Correlation is not causation
First, the NIEHS scientists discussed the comparison of lifetime risk of certain site-specific cancers with total stem cell divisions in the associated tissues. Statisticians use a value called R squared to represent the correlation between two different measured variables. The closer the value is to 1, the more closely correlated the two variables are.
Tomasetti and Vogelstein reported the R squared value for their data as 0.65, concluding that 65 percent, or about two-thirds, of the differences in cancer risk among different tissues can be explained by the total number of stem cell divisions in those tissues.
And this is where terminology becomes important. “’Explaining’ is statistical jargon that has little to do with causation — it has to do with the relationship between X and Y,” Weinberg said. “You can’t conclude from an R squared value how much of that causal relationship is actually due to the variable that X represents.”
Aggregate risk and multiple causes
A second concern Weinberg and Zaykin raised was the use of aggregated lifetime risk data, a point raised by other commenters. “The R squared from an analysis of cancer types based on aggregated risk for each type obscures the contributions of individual risk factors to each cancer type,” they wrote.
Finally, they addressed the researchers’ approach of partitioning different causes into fractions that add up to 1. Referring to phenylketonuria, an inherited condition in which the body cannot properly metabolize phenylalanine, they wrote, “The fraction attributable to genetics is 1.0, while the fraction attributable to environment is also 1.0, because the outcome requires both a dysfunctional metabolic gene and an environmental exposure [dietary phenylalanine].”
“Weinberg and Zaykin make important points,” said NIEHS Scientific Director Darryl Zeldin, M.D. “The claim that two-thirds of the cancers studied are caused by bad luck could lead to an overemphasis on development of treatments, at the expense of crucial research into prevention and the role that environmental and genetic factors play in the origins of cancer.”
Other commentary, other forums
Weinberg and Zaykin were not the only NIEHS scientists to enter the debate. NIEHS biostatistician Shyamal Peddada, Ph.D., posted an online comment on the Science paper, suggesting that the authors’ mathematical approach, which used log-transformed data, led to a faulty conclusion. “[It] is not supported by the raw data given in their Table S1,” Peddada wrote. “A correlation analysis of the raw data reveals a weak correlation, with only 28 percent of the variability in lifetime risk of developing cancer explained by the total number of stem cell divisions.”
Besides commentary on the Science website, the paper was vigorously discussed in PubMedCommons (see sidebar). The public debate underscores the important contribution that open dialogue and data access can make to ensuring that new scientific findings are properly interpreted, so they can be confirmed and appropriately followed up in future studies.
Tomasetti C, Vogelstein B. 2015. Cancer etiology. Variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science 347(6217):78-81.
Weinberg CR, Zaykin D. 2015. Is bad luck the main cause of cancer? J Natl Cancer Inst. 107(7):djv125.