Most anyone seeing the sign above reads it as “PARIS IN THE SPRING.” In actual fact, the word THE is repeated twice so that the sign actually reads “PARIS IN THE THE SPRING.” Your mind was acquainted with the phrase so that it “saw” what it expected. This is just one example of the problems that a human observer has in seeing what is actually present in an image. It is the reason that we have spell checkers in computers and the reason that having more than one radiologist look at an imaging study (double reading) can reduce the false-negative rate for any search.

The vagaries of perception are clear to anyone who has ever looked for his or her keys in the morning. You are very clear on what your keys look like, but as hard as you look you do not see them until someone else points them out to you in plain sight on the table. The fact that you did not see the keys does not make you negligent or an evil person (as the US medicolegal system tries repeatedly to prove), but merely reflects the basic problems inherent in human perception. The ability to see something is a complex psychovisual process that is clearly far from perfect. For whatever reason, we all (without exception) periodically fail to see obvious things no matter how hard we try. This was evident in the kindergarten task of trying to find all of the animals hidden in the barnyard picture. It is the reason for the success of the “Where's Waldo” cartoons.

The effort to perceive information in a distracting background is part of the fun of reading the Arts & Leisure section of the Sunday New York Times. The artist Hershfeld draws the name of his daughter Nina into his caricatures of famous people from the entertainment world.<sup>11 Kopans D.B.   Breast Imaging Philadelphia: Lippincott-Raven (1998). 

Cliquez ici pour aller à la section Références</sup> Hershfeld is kind enough to indicate how many Nina's there are in his drawings by writing the number next to his signature. My son and I used to see who could find the names in the drawings. In one memorable drawing, according to the key, there were five Nina's. We each found three of the names. He found one that I missed and I found one that he did not see. That is an example of double reading. We both were unable to find the fifth Nina. That is a lawsuit!

One can imagine that the task is far more difficult if the observer does not know the name that is hidden in the drawings, and does not know which of the drawings contains a name. The process seems almost impossible if there are only 10 images out of 1000 presented to the observer that contain a name, yet this is exactly the task facing the radiologist who interprets mammograms. These puzzles and our ability to appreciate (and often fail to appreciate) the hidden information, no matter how skilled we may be in observation, are all examples of the vagaries of human vision and perception. Given that there is not a radiologist in the world who has not overlooked a significant finding on a radiograph or a cancer on mammogram that is visible in retrospect, it is baffling that any radiologist could testify that another was negligent in not seeing a cancer that is visible in retrospect.