Sir Godfrey N. Hounsfield, who won the 1979 Nobel Prize in Medicine for inventing computed axial tomography, died Aug. 12. He was 84.
The first CT machine took nine days to scan an object. Processing the information by computer then took 21 hours. Despite such formidable obstacles, researchers were undaunted.
Hounsfield conceived the idea for a CAT scanner in 1967. By 1972, he had built a machine that produced detailed cross-sectional images of the brain in less than five minutes. He began work at EMI in Middlesex, U.K., in 1951 and retired in 1986.
Introducing Hounsfield at the Nobel presentations, Prof. Torgny Greitz of the Karolinska Medico-Chirurgical Institute said that the first clinical results in the spring of 1972 flabbergasted the world.
"Up to that time, ordinary x-ray examinations of the head had shown the skull bones, but the brain had remained a gray, undifferentiated fog. Now, suddenly, the fog had cleared," Grietz said.
Hounsfield's own words best describe the genesis of CT scanning. The following text is excerpted from his 1979 Nobel speech:
"When I investigated the advantages [of CT scanning] over conventional x-ray techniques, it became apparent that the conventional methods were not making full use of all the information the x-rays could give.
"Calculations showed that computed tomography used the data very efficiently and would be two orders of magnitude more sensitive than conventional x-rays. For this reason, I hoped that it would be possible to distinguish between the various tissues of the body, although I could not find any literature which suggested that such x-ray absorption differences existed.
"I decided to do some lab experiments with gamma rays to test if the system would work. The equipment was very much improvised. A lathe bed provided the lateral scanning movement of the gamma ray source, and sensitive detectors were placed on either side of the object to be viewed which was rotated one inch at the end of each sweep.
"The 28,000 measurements from the detector were digitized and automatically recorded on paper tape. After the scan had been completed, this was fed into the computer and processed.
"Many tests were made on this machine, and the pictures were encouraging despite the fact that the machine worked extremely slowly, taking nine days to scan the object because of the low intensity gamma source. The pictures took 21 hours to be processed on a large computer.
"Clearly, nine days for a picture was too time-consuming, and the gamma source was replaced by a more powerful x-ray tube source, which reduced the scanning time to nine hours. We produced the first picture of a brain [specimen] to show gray and white matter.
"In parallel, tests were carried out on sections through pigs in the area of the kidneys, and this work also produced most encouraging results. Although the speed had been increased to one picture per day, we had a little trouble with the specimen decaying while the picture was being taken, so producing gas bubbles, which increased in size as the scanning proceeded.
"We found that we could see brain and body tissues clearly, but we were still very worried as to whether tumors would show up at all. Unless it could do this, the machine would be of very little use. To test this, we had to build a much faster and more sophisticated machine that would scan the brains of living patients in a hospital.
"In 1972, the first patient was scanned by this machine. She was a woman who had a suspected brain lesion, and the picture showed clearly in detail a dark circular cyst in the brain.
"Applying the principles to scanning the body seemed to be the next logical step. A larger and faster scanner was designed capable of taking high-resolution pictures of the body in 18 seconds.
"What improvements should we expect to see in the future? Various attempts have been made to achieve useful pictures of the heart. [One] approach is being made at the Mayo clinic, where a large machine is being constructed with 27 x-ray tubes designed to fire sequentially. It is hoped to take a sequence of pictures in a fraction of a second during one heartbeat. However, the complexity and cost may rule out such a machine being used worldwide.
"A further promising field may be the detection of the coronary arteries. It may be possible to detect these under special conditions of scanning.
"As all the information on the body is stored in three dimensions, it is possible therefore to display the object at any angle. This allows it to be examined by rotating it around on the screen. The views seen around the organ to be examined may reveal information that hitherto could have been missed, when it was viewed normally in one fixed plane, normal to the axis of the body."
For more information from the Diagnostic Imaging archives: