Practical evaluation of image intensifier systems. "Information" vs. dose rate.

Abstract

The wide application of image intensifiers in clinical fluoroscopy and cineradiography has raised the problem of the evaluation of the various systems. Much emphasis has been placed on the comparative intensification of light output and this emphasis has become more evident with the adaptation of closed circuit television systems to the image intensifier tube, making attainable brightness intensification of over 50,000 times. The importance of light intensification should not be minimized, but a more important characteristic, the ability of the system to convey “information” at a given x-ray dose rate, has received less attention. A bright output phosphor or television monitor is of poor diagnostic value if insufficient detail and contrast are present in the image. The ability to increase the light output of the intensification systems far exceeds their ability to produce a comparable increase in image detail. The degree of light intensification produced by optimal systems can approach 50,000 to 60,000× as compared to the standard Patterson B-2 screen. However, the increase in perceptibility of detail is not proportional to the greater brilliance of the image. A more practical criterion of the performance of an intensification system is the intensity of x-radiation at the input surface (r per minute) of the intensifier tube necessary to produce an image of given detail and contrast with a standard test object. This value constitutes a figure of merit in the intercomparison of various intensification systems. A given image is composed of patterned “bits of information,” i.e., quanta of energy, in a background of similar particles randomly distributed. Identification of the image is dependent on separation of these bits of information (signal) from the background (noise). This is usually achieved by having many “bits” (a strong signal) in a weak background (low noise). Increasing the signal to accomplish this, however, necessitates the raising of the x-ray dose rate. An ideal image intensifier tube utilizes a low dose rate by intensifying the signal and noise, while maintaining their differential ratio. The dose rate may be reduced to the point where the signal is lost in the noise and, therefore, a limit is reached. The addition of television to an image intensifier can extend this limit by means of contrast control. This is dependent on the fact that, as contrast is increased, detail perceptibility improves. However, contrast enhancement also increases the noise of the system, and the resultant image is “snowy.” We have tested various intensifier systems for detail rendition with either wire mesh screens or thread gauges, determining resolution in “lines” per inch as with a photographic test pattern. A 9-inch Philips image intensifier with a mirror optical image system (“carrousel”) will resolve 60 lines per inch. This technic is unsatisfactory in that detail perception of this kind tests only one aspect of image formation.