Computed Tomography — Past, Present and Future

Abstract

The basic principles of computed tomography go back to the work of J.H. Radon, a Bohemian mathematician, who in 1917 published the mathematical framework for reconstruction of an object from its line integrals [1]. First experiments on medical applications of tomography were conducted by A.M. Cormack [2], a hospital physicist at Groote Schuur Hospital in Kapstadt, who without knowledge of Radon’s prior work, set out to improve radiation therapy planning. Cormack developed a method for reconstructing the absorption coefficient of a slice of the human body from transmission measurements but was not able to prove the medical significance of this invention. Only later, in the 1970s, did Cormack learn about the prior work of Radon and first applications of his theory in radioastronomy [3]. G.N. Hounsfield, who today is widely recognized as the inventor of computed tomography, independently discovered the method in 1972, and was the first to develop a successful practical implementation [4]. He began his experiments using radioisotopes as sources for his transmission measurements, with measurement times on the order of 9 days. Using more powerful X-raytubes, it still took about 9 h to complete a measurement. Nevertheless, Hounsfield was able to install a first prototype of his CT system at Atkinson Morley’s Hospital in London, and, working closely with neuroradiologist J. Ambrose, successfully scanned the first patient in 1972. A photograph of the early system is shown in Fig. 1. This development sparked a wave of excitement in the medical community. For the first time it was possible to obtain cross-sectional images of the head free of superposition with dramatically improved low-contrast resolution capabilities. In 1979 Hounsfield and Cormack were awarded the Nobel Prize for their invention.