For instance, quantitative comparison of the activity uptake of left and right organs or lobes is possible. Figure 12.5 demonstrates how we measure relative kidney function.

12.4.4.2 Information Extraction

A dynamic image series may contain a huge amount of information. During a hepa­tobiliary study, for example, 110—120 images are typically taken, each represented by a 128 X 128 matrix—a matter of nearly 2 million numbers that is impossible to analyze entirely by the human eyes. We need computers to extract useful infor­mation from the raw data that can be related to the function of the investigated organ. Figure 12.6 demonstrates the possible solutions:

• By drawing the contours of the organs (ROI = “region of interest”), time—activity curves from the counts inside these regions are generated, and some parameters are calculated from them.

• Alternatively, we may calculate a parameter from the time—activity curve of each pixel, reinsert its value to the respective element of a matrix, and display the resulting “parametric image” as a pseudocolor image.

The two methods can also be combined: parametric images may help with outlining the appropriate regions.

Figure 12.5 Split renal function calculated from the geometric mean of the posterior (PA) and anterior (AP) images (left: 14%, right: 86%).

Parametric images

Figure 12.6 Methods for processing a dynamic image series. (Images of the kidneys are shown.)

12.4.4.3 Archiving

Traditionally, X-ray films were used to display and archive medical images; however, this method is rather expensive. Nowadays, it is cheaper by two orders of magnitude to archive the image series of nuclear medicine on CD or DVD.