Much ado about (almost) nothing

In the June Lancet Dr Mark Pearce, et al. published an article, "Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumors: a retrospective cohort study." full article and a comment.

In the November issue of Radiology, Drs Brenner and Hall, both radiation physicists of some renown, published an editorial referring to this article entitled, "Cancer Risks from CT scans: Now that we have the data, what next?" full editorial

It is a catchy title for an editorial, but the problem is we still don't have the data.  The study by Pearce, et. al. compares the long term followup of children who have no CT scans vs kids with a low number of CT scans vs kids who have several CT scans, and looks at the incidence of brain tumors and leukaemia.  It is a large study with some 180,000 children included.  What is not included is why they got imaged in the first place, as this could not be extracted from the database.  They conclude, and of course Drs Brenner and Hall agree, that the CT scans caused the tumors.  What they actually proved is that kids who require multiple CT scans are sicker kids. 

Every physician with any experience knows that some people are just more disease prone.  In the old days of film radiology, these were the patients with the thick jackets.  Now they are the patients with the long problem lists in their electronic medical record. When I was an intern we called it PPP (piss poor protoplasm.)

The study retrieved data from national registries on children who had CT scans before the age of 22.  They then retrieved data from cancer registries for two diagnosis, leukemia and brain tumors.  They did not look at other diagnosis.

I would be willing to bet there is a long list of diagnosis which are statistically different in the three groups.  They need to analyze the data for the incidence of things like intestinal disorders, pneumonia and chronic pain in these children down the road.  No doubt these problems are more frequent as well.  Did radiation cause them all?  They want to believe the CT scans caused the cancer, but they have used selection bias to divide their study groups.

Just another example of researchers wanting to support their preconceived beliefs.