Cancer By the Numbers

There is always a novel attempt to garner information or patterns from data on cancer. In a recent Science paper Vogelstein and colleagues have done some interesting "back of the envelope" analysis. They did the following:

1. Collected data which provided the number of stem cells of a particular cells type in a typical lifetime of a person. For example they somehow determined from the literature the number of melanocyte stem cells in a lifetime. They did the same for many other cells. For example there are lots of basal cells per human lifetime, as one would expect. In contrast there were two orders of magnitude less melanocytes.

Let us call that NSC(i) where i equals a specific type of cell.

2. Then they plotted the incidence of cell related cancers versus the total lifetime stem cells by type. Thus we see that the incidence of basal cell cancer is high and so are the total number of lifetime stem cells of basal cells.

Let is call this INC(i) the incidence of a specific cancer related to the specific cell.

3. Then they created a normalized line through the incidence versus stem cell count curve and drew a chart of cancers how far below to how far above they were to the average line. This waterfall type chart then was the discussion point.

That is we have some generalized relationship:

INC(i)=K NSC(i)  where K is a common constant obtained from a regression type analysis. However the actual INC(i) may be above or below the regression line.

4. The cancers well above the norm were those driven by some putative genetic or environmental factor such as smoking and lung cancer. The rest are due the authors state to just having lots of stem cell mutations.

This if INC actual (i) > INC regression (i) we attribute this excess to some genetic or environmental/lifestyle condition.

Interesting concept, but there are some issues:

1. Do they really mean stem cells? Melanocytes do not reproduce as quickly but it is not clear just what a melanocyte stem cell is. We have seen this in prostate cancers.

2. The authors admit epigenetic factors as well and one suspects that they could dominate.

3. The excess cancers such as smoking and lung are clearly environmental effects.

4. Some how there is no discussion of breast and prostate. One wonders why since they are so prevalent.

Otherwise this is interesting and worth the read. Their conclusion is:

These results suggest that only a third of the variation in cancer risk among tissues is attributable to environmental factors or inherited predispositions. The majority is due to “bad luck,” that is, random mutations arising during DNA replication in normal, noncancerous stem cells. This is important not only for understanding the disease but also for designing strategies to limit the mortality it causes. 

That is worth exploring.But, and this is a classic case, the Press has latched on to the "bad luck" phrase. We really do not know from this study what is the issue. Thus again we have a confluence of words by authors and the explosion of the press to enhance the piece. Frankly it is interesting but hardly conclusive of anything!