FIGURE 16 Implementation of standardization functions in clinical PACS.

case there is a failure of preprocessing functions. With standardization performed on acquisition gateway, if undesirable modification of the image occurs, the radiological procedure has to be repeated. A comparison of both installations is made in Table 1.

The standardization function to be implemented depends on the modality as well as the anatomy shown in the image. Not all of the images have to be subjected to all preprocessing functions discussed in this chapter, including correction of image format, background, and orientation, search for window/level values, optimization of brightness/contrast, or correction of image size in order to improve the accuracy of quantitative measurements. Thus, a set of functions performed on a certain image is precisely defined for each type of image. Depending on the application, preprocessing functions can be divided into two categories: only modality-dependent and both modality- and anatomy-dependent.

Clinical experience has shown that each modality presents unique problems that need to be resolved before soft-copy images are suitable for diagnosis. In this chapter, preprocessing techniques for four modalities are considered: angiography, CR, CT, and MR.

Quantitative analysis and measurement of recorded blood vessel diameters play an important role in angiography. It serves as a basis for other values or indexes to be derived. This requires implementation of a technique that reduces distortions caused by mapping a flat surface onto a spherical image-intensifier surface. Functions suppressing those distortions are discussed in Section 5 of this chapter. Since no danger of image destruction has been observed, this function can be implemented at the acquisition gateway.

CR images require preprocessing functions that remove the unexposed background (Section 2.1), standardize the orientation (Section 4), and improve the visual perception by adjusting selected anatomical regions (Section 3). Background removal may not be applied in adult chest or abdomen, yet it should by performed in hand, limb, and head images. On the other hand, about 70% of pediatric images show a wide unexposed background, which needs removal. Therefore, all pediatric images are subjected to a correction procedure. Since the methodology does not depend on the anatomical structure within the image (only background is subjected to the analysis) one function handles all anatomies.

Although this function is installed at the acquisition gateway, a possible image damage issue needs to be considered carefully.

The procedure discussed in Section 2.1 has been clinically tested and no removal of the diagnostic field of view has been reported. Yet, the function reaches a certain level of accuracy in a particular clinical environment and technical setup, including radiological equipment, type of collimators, sensitivity of image plates, etc. Thus, before being installed in another environment in an unsupervised mode, the function should be tested carefully on a large set of clinical data.

Since about 35 to 40% of images acquired in radiology are not in an upright position, standardization of image orientation becomes an important issue. Requiring manual image rotation with a selection of the rotation angle would interrupt the medical diagnosis based on soft-copy display, as well as image presentation during seminars, conferences, and teaching. Methods for orientation correction discussed in Section 4 of this chapter do not destroy the image content. They may place the image in the wrong orientation, but this failure is reversible. Thus, the function should be installed at the acquisition gateway. Moreover, if rotation is required, the entire procedure becomes time-consuming. Its implementation on the archive gateway or workstation would significantly lengthen the overall system response time. The delay depends on the image size.

For CR as well as CT/MR images, certain LUTs are required to enhance one anatomical region and suppress the others. Fixed LUTs are of particular importance at viewing stations for a preview of images, during conferences, teleconferences, and teaching. They also are used during a primary diagnosis to enhance various gray-scale regions (i.e., lungs and mediastinum in chest images) at different levels. Thus, the best place for the implementation of those preprocessing functions is the acquisition gateway. At this time enhancement parameters are stored in the image header. Since no penetration into the image content is performed, no damage can be done to the image information.

0 0

Post a comment