Image display has traditionally been implicated as the limiting factor in the usefulness of digital data processing in nuclear medicine. This was partly due to limited technology and partly to an attempt to use a single system for all displays. Because of the widely varying characteristics of the various nuclear medicine procedures, no single system can perform optimally in all studies. The various uses for the display in data processing are: user interaction, feature extraction, quantitative parameter estimation, correlation with previous studies and reporting. User interaction poses the least stringent requirements on the display. Typewriters, line printers, and standard computer terminals are adequate for this mode of operation. Display of monitor images to allow the operator to follow the processing of the data requires only a moderate quality display, and, for this, a storage scope display or gray scale with a limited number of gray scales is useful. Interactive displays, which are used to guide the processing, must have some means of allowing the user to indicate regions of interest. Joysticks, cursors, or graphic tablets can perform this function. For display of images to be used for diagnostic interpretation, a high-quality image must be produced, which allows the entire dynamic range of data to be shown to the clinician. Printing of multiple images on an electrostatic plotter or on film used in a flying-spot scanner mode can provide the clinician with such images. Display of dynamic data places another restriction on the system in that not only must spatial distribution be represented, but the temporal distribution of activity must also be shown. This is approached in one of three ways. If the problem is one of changing organ size, as in a cardiac ventricular ejection fraction measurement, then the overlaying of the organ edges as a function of time, along with a curve of the change in the area (or counts) as a function of time, is adequate. If, however, the problem is one of measuring a single organ function as a function of time as in a cardiac bolus study where the presence of a shunt is suspected, then one can use time-coded color displays in which the activity distribution in selected time segments is displayed in different colors. In this case, the overlapping activity areas are shown as the summation of the individual colors, while the nonoverlapping areas remain in their respective colors.
Read full abstract