What is normal?
hen Warren G. Harding became President of the United States in 1920, shortly after the First World War, his motto was Back to Normalcy. Unfortunately, Harding and his political friends never gave their definition of normalcy – their approach to normalcy in the domestic politics of the U.S.A. was rather stunning and dreadful.
In medicine, the concept of normalcy is different, but also rather unclear. In the early days of roentgenology, two standard books in radiology were published by German professors. The first one was written by Rudolf Grashey in 1905 , the second one by Alban Köhler in 1910 . Since then, numerous reprints and new editions have described the borderlands of the normal range and the beginning of pathology in x-rays, and today there are many other books on the same topic.
Over the decades, tens of thousands of x-ray images of every part of the body have been taken to create a catalog of normal features and of varieties of the normal. The result is an overview of normalcy and the delineation of the borderlines to pathology.
With the appearance of MR imaging in routine clinical practice, an abundance of new insights arose in clinical imaging. Radiologists were confronted with tissues and tissue changes that previously had only been accessible to, and known by, pathologists. For instance, nobody in clinical diagnostic medicine had ever seen such accurate and distinct slices of the brain as MR imaging could now produce. Physicians have had to relearn anatomy and pathology.
Therefore, MR imaging has been a great boost for publishers of anatomy books, and the market for comparative books and CD-ROMs of anatomy with imaging techniques is still surging.
As new structures became visible, image reading was transformed into an even more delicate and difficult task, and contrast behavior was unpredictable, given the multitude of parameters influencing image contrast. Once again, the borderlines of medical normalcy and its variations, which should not be assessed as disease, were unclear.
The diagnosis of multiple sclerosis is the standard example of what can happen if there is no proper knowledge of the normal range. Multiple sclerosis plaques are easily visible in MR imaging, and thus the possibility of verifying the diagnosis proved attractive to an enormous number of physicians, as well as patients and their relatives.
MR imaging revealed white-matter lesions in many patients, who were therefore diagnosed as having multiple sclerosis. However, soon it became evident that such white-matter lesions could also be observed in control groups of normal volunteers. High-signal-intensity spots, called unidentified bright objects (UBOs) by some authors, were detected in both healthy subjects and patients with a variety of diseases or conditions .
The appearance of these spots is consistent with an increased water content and changes in myelin structure. A local lesion (e.g., a cerebral edema caused by circulatory changes or breakdown of the blood-brain barrier) may result in atrophic perivascular demyelination, myelin pallor, gliosis, infarction, and/or porencephalic changes, all of which can be seen as hyperintense spots in either intermediately or T2-weighted MR images.
Initially, these spots created some confusion, but soon this was removed. Joseph Durand, a French physician from Lyons, had described such lesions already in 1843 and given them the technical terms état lacunaire and état criblé. Now they could be seen in vivo in patients.
Several studies showed that the finding of UBOs was common in MR imaging, but that such changes are unusual in individuals less than 40 years of age. However, the frequency increases with age. It was also found that risk factors for cerebrovascular disease and a history of brain ischemia correlated positively with the number of lesions. Smoking, a known risk factor for atherosclerosis, correlates with the increased occurrence of changes . However, many of these patients and volunteers had no neurological symptoms or psychometric changes. From a health point of view, they were normal.
Whereas previously it would be usual to read about findings consistent with demyelinating disease, today such descriptions are (hopefully) worded far more carefully, and they are only found if the clinical history suggests multiple sclerosis or another of the possible causes for white-matter lesions. In some reports, single bright spots in the white matter of the brain are still described in the findings section, but there will not be a pathology description in the impression section of such reports. The spots are not mentioned because they are not considered to be of clinical relevance, and the verdict is normal.
Then the question arises as to where is the border between normalcy and pathology and when such spots should be considered pathological. This question can become extremely important if there are far-reaching consequences for the patient or, as in the following case, for a group of people who might be potential patients.
As part of a larger study, the brains of a group of deep-sea divers were examined. These divers stay and work at great depths below sea level for several weeks. Before they can return to the surface, they have to undergo decompression. It has been postulated that decompression can induce minimal brain damage, which over the years leads to permanent damage.
As with all such studies, one needs a reference group with which to compare the results of the target population. This reference group should be similar but normal. In the case of the divers, police officers and off-shore workers were chosen as occupational groups having to fulfil the same stringent medical selection criteria; but they were not diving.
The MR imaging results were striking. The white-matter changes for the divers and the control group were 33% and 43% respectively . According to the literature, such changes should be expected in no more than 20% of the population.
Obviously, the control group did not resemble the normal population, but to make sure, another randomly chosen group was examined and they revealed less than 20% of such changes, as expected. It is unclear why the offshore workers/police group showed more white-matter changes than the divers and the normal population.
However, in epidemiological terms, choosing them as a control group was wrong.
What are the lessons of this study?
First, despite the fact that millions of MR imaging brain examinations have been performed all over the world during the last two decades, normalcy ranges still have not been set. This is true not only for the brain but also for the spine, liver, etc.
Second, selecting a control group for clinical studies may be a more difficult task than is generally thought, particularly if such a group’s normalcy has not been defined. This means that the results of such studies may have to be interpreted cum grano salis.
Some years later, functional MR imaging of the brain became possible and fashionable. Again, the question has to be asked: what is normal and was is pathological – and what is an artifact?
Many blood flow alterations described in functional brain imaging rely on signal-intensity changes of less than 5%.
As Gustav von Schulthess once pointed out:
“... a caveat for fMRI: it is a very interesting technique but signal changes are but a few percent. Hence, the method is technically demanding and ‘the threshold of nonsense production is low’ .”
1. Grashey R. Atlas typischer Röntgenbilder vom normalen Menschen. Munich: Lehmann 1905.
2. Köhler A. Grenzen des Normalen und Anfänge des Pathologischen im Röntgenbilde. Hamburg: Gräfe und Sillem 1910.
3. Refer to list of references cited in the article in reference 4.
4. Rinck PA, Svihus P, De Francisco P. MR imaging of the central nervous system in divers. J Magn Reson Imaging 1991; 1: 293-299.
5. Todnem K, Skeidsvoll H, Svihus R, et al. Electroencephalography, evoked potentials and MRI brain scans in saturation divers. An epidemiological study. Electroencephalogr Clin Neuro physiol 1991; 79: 322-329.
6. von Schulthess G. Clinical MR in the year 2010. Mag Res Med 1999; 8: 133-145