he combination of MR imaging with extrinsic contrast-changing agents led many researchers to try to highlight anatomical and pathological structures and even metabolic processes which, per se, are invisible on single plain images [⇒ Rinck]. Image series of the same anatomical structure during a time period before and after the application of a contrast agent add another dimension, commonly described as perfusion imaging. Perfusion describes blood delivery to tissues, usually flow on the capillary level.
Perfusion imaging can be categorized into two classes: those methods monitoring target-tissue signal changes after the application of an extrinsic contrast agent, and those relying upon intrinsic factors such as increased blood volume or blood oxygenation and flow in the microvasculature. The former can be visualized by dynamic imaging, the latter by functional imaging.
However, even the individual images of a dynamic time series will, in some cases, not yield all information contained in them. Thus, the combination of dynamic MR and mathematical image manipulation has been proposed (Figure 16-01).
This chapter gives an overview of some of the techniques. It cannot be exhaustive considering the wide spectrum of the field. For detailed scientific treatises see Maintz and Torheim [⇒ Maintz, ⇒ Torheim 1999].
Traditionally, the following methods have been applied to analyze dynamic contrast-enhanced images:
visual inspection of time-intensity curves;
visual inspection of the time series by running the images in a movie;
Yet, the most intriguing and interesting way of processing the data is by the creation and subsequent visualization of parametric maps: images combining parametric images derived from the information present in the image series with anatomical information (Table 16-01).
To be clinically useful, such a postprocessing method must be robust, reliable and automatic or semi-automatic.
Steps of the entire dynamic acquisition and image-processing procedure. The noise-filtering and motion-correction steps are optional, but in many instances necessary and fairly complicated.
16-02 Inherent Problems
In order to process the time series (= the series of images starting with one or several precontrast images followed by the contrast-enhanced images) on a pixel-by-pixel basis, the spatial location of picture elements within the image matrix must be exactly the same. In many instances, this is not the case due to movement (Figures 16-02 and 16-03).
Four images taken out of a dynamic study of the kidneys. During the examination, both kidneys move in all three dimensions and rotate. In this example, the right kidney moves several pixels up and down during the time series (red line). Mathematical processing of the images can help eliminate some of the movement and facilitate processing the time series.
All dynamic images in this chapter were produced with Dynalize 1.0 [⇒ Torheim 1997].
This hampers the calculation and evaluation of time intensity curves and parametric images [⇒ Gehrig]. Pixels or regions-of-interest have to be reallocated manually or, preferably, automatically, e.g., by contour recognition. The realignment of the images is often referred to as image registration.
This problem is even more complicated in contracting organs such as the heart [⇒ Higgins]. Here artifacts can be corrected by highlighting the edges of the organ by drawing the boundaries either manually, semiautomatically or automatically if there is some a priori knowledge, e.g., previous knowledge of the outline of the organ and the ratio between its length and width (Figure 16-04) [⇒ von Schulthess]. A more sophisticated approach is the automatic creation of regions of interest (auto-ROIs) [⇒ Torheim 1997].
Left: A reference image is selected.
Center: The boundaries of the kindey can be manually drawn and adjusted in each image of the examinations series.
Right: An edge-detection program can define the boundaries through contour-enhancement and automatically or semiautomatically adjust them by comparing the ratios between fixed parameters.
Image registration of contrast-enhanced images of non-rigid organs is a difficult problem due to the changes not only in contrast but also in the structures visible. Whereas rigid organs like the brain can be aligned by performing translations and rotations, non-rigid organs require non-rigid reformations, which potentially can change the structures to be observed.
An additional problem is the propagation of mistakes, e.g., through artifacts which are not recognizable any more on processed images. They can be caused by a change of relative or absolute signal intensity on the image influenced by outside factors such as surface coils.
Furthermore, due to the frequent presence of time constraints, the signal- to-noise ratio is often quite low in dynamic contrast-enhanced MR imaging. This necessitates the need for noise-reduction techniques [⇒ Sebastiani] which are commonly performed by applying so-called filters either in the spatial (image) or in the temporal domain.
There are some very sophisticated filters which can remove noise in a way which was previously only possible by increasing the magnetic field strength or improvements in the amplifiers or other hardware.