Sunday, December 16, 2007


I wrote about the epidemiology of myelodysplastic syndrome (MDS) five years ago and I have now been asked to update the chapter. The only big change has been the publication of SEER figures. The Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute (NCI) is an authoritative source of information on cancer incidence and survival in the United States. SEER currently collects and publishes cancer incidence and survival data from 17 population-based cancer registries covering approximately 26 percent of the US population. The SEER population groups closely mirror the US population mix, though rural populations are slightly under represented and non-white populations slightly over represented.

SEER suggests that the incidence of MDS is 3.6 per 100,000 which is exactly the same figure that the British LRF study came up with some 15 years ago (this sampled 16 million of the British population), and the same as the German Dusseldorf study (looking closely at 1.2 million) concluded 10 years ago. So why do I doubt it?

The first reason is that I have had experience at diagnosing MDS and I know it isn't easy. In MDS, the abnormal cells do not come labelled, they have to be recognized through rather subtle changes. The neutrophils don't have so many granules in their cytoplasm and their nuclei don't have so many lobules. You get a ring of iron (or more commonly part of a ring) around the nucleus of the erythroblasts in the bone marrow - but this doesn't show up with normal stains - you have to stain for it in a special way. The megakaryocytes are too small or not lobulated properly. There are sometimes too many blasts in the bone marrow - but the blasts often don't look quite like blasts, more like promyelocytes. Just like CLL the patients are usually asymptomatic when they present and they are spotted because someone does a blood test. Often the blood count figures are normal, so it takes a close look at the blood film by an expert to make the diagnosis, and even then a bone marrow may be necessary to be sure. Until I learned how to recognize MDS in 1981, I had missed 76 cases in the past 5 years.

The second reason is that cancer statistics draw on cancer registries, and cancer registries rely on hospital admissions and death certificates for their figures. Since MDS rarely gets mentioned on death certificates and many patients are never admitted to hospital, the diagnosis is never recorded in many cases. The LRF study went to the only group of people who could accurately diagnose MDS, the haematologists. They were asked to record every case that they diagnosed over a five year period. The figure of 3.6 per 100,000 excluded patients over-80 years of age - the very people in whom it is commonest. When these were included in a later publication the incidence rose to nearly double what had been previously thought.

The third reason that I doubt the figures is a four-fold variation between different towns and cities. The highest incidence found was in our own laboratory at 12.7 per 100,000. The main factor that accounted for this difference was how interested the local hematologist was in the disease.

Even 12 per 100,000 is an underestimate of the true incidence because when we set about screening for the disease in people over 50, by undertaking an unnecessary blood count we found a prevalence of about 1 in 500! There is a lot more out there than one might expect.

Of course it doesn't really matter, because most of these cases will never be of clinical significance - just as cases of MGUS or monoclonal B-cell lymphocytosis will never draw attention to themselves. Almost certainly most cases of prostate cancer picked up by PSA screening and many cases of breast cancer picked up by mammography would never have troubled the surgeon.

No comments: