The Maximum Intensity Projection

Maximum intensity projection (MIP) post-processing is the most common means of displaying MRA data as the views are comparable to those yielded by conventional x-ray angiography. To obtain a MIP reconstruction, parallel rays are cast through the 3D data set (i.e. volume) with each pixel of the projection image plane representing the maximum signal intensity encountered along the corresponding ray. With this technique, a ray is projectedthrough the 3D data set in the desired direction, and the highest voxel value along the ray becomes the pixel value of the two-dimensional MIP image. Thus, the MIP algorithm does not simply add up the signal intensities from slice to slice along the projected axis. Instead, it finds the highest signal intensity value (i.e. maximum intensity) along a projection line. In this way 3D data can be collapsed into a 2D projection image. MIP process ing of 3D CE MRA data permits clear visualization of vessel enhancement, producing views that closely resemble those of conventional x-ray angiogra-phy. For this reason MIP reconstructions tend to be clinically popular for viewing and interpretation of 3D CE MRA. This method is particularly well suited for image interpretation of arterial-phase CE MRA, in which the arterial-to-back-ground signal difference (i.e. contrast-to-noise ratio or CNR) is high. Most MR scanner manufacturers now include options for automated MIP processing of data in pre-determined orientations (e.g. axial, sagittal and coronal) at the end of the image acquisition process. This allows a rapid assessment of the vessels while the patient is still in the scanner.

Unfortunately, some disadvantages are associated with automated MIP algorithms that typically include all data within the entire volume (Fig. 3). First, full volume MIP reconstructions tend to misrepresent anatomic spatial relationships since depth information is not displayed [16]. Eccentrically located stenoses may remain undetected and superimposition of structures may erroneously simulate the presence of a stenosis [17]. Second, full volume MIP reconstructions tend to increase the mean signal intensity of the background [18] [19]. Consequently, vessels with low signal intensity, which may be seen on individual source images, may be partially or completely imperceptible on full volume MIP images [20]. Since the MIP algorithm does not differentiate the etiology of signal intensity, it may not project a vessel sufficiently if its signal intensity is equivalent to that of its background. Regions of high signal intensity are invariably seen on CE MRA. Since the pulse sequence is T1-weighted, tissue with short T1 times such as fat or bone marrow might degrade MIP viewing of CE MRA acquisitions. The likelihood of increasing background signal (i.e. maximum intensity signal along the ray in non-vascular tissue), increases

High Thrombus

Fig. 3a-c. MIP of lower limb arteries (a), corresponding coronal "source image" (b) and axial MPR (c). The dark thrombus (arrows) in the left popliteal artery aneurysm is well visualized on the coronal source image and the axial MPR image, but is not appreciated on the coronal MIP image with the inclusion of additional non-vascular regions. Thus, a vessel which may be seen on an individual source image may not be apparent on a full volume MIP reconstruction if there is a lot of spurious high signal tissue overlying the region (e.g. subcutaneous fat above and behind the vessel). Furthermore, full volume MIP images obtained from the entire 3D data set are frequently contaminated by wraparound or edge artifacts, whichcan limit the visibility of vessels. On full volume MIP reconstructions, overlapping vessels, whether arterial or venous, may mask an arterial stenosis or occlusion thereby making evaluation impossible. Finally, depth information is not represented on MIP reconstructions. Therefore, it is often difficult to tell the relationship between stray intensities and blood vessel or to separate the signals from two overlapping vessels. In order to compensate for the loss of the depth information on standard full volume MIP images, post-processing systems provide such solutions as: 1) enabling MIP reconstructions of a sub-volume of interest, instead of projecting the whole volume, and 2) varying the projection angle (interactive rotation).

Although the generation of rotational sub-volume MIP images is often routinely performed by the technologist at many sites, this alone is rarely sufficient for final diagnosis. Therefore more intricate additional post-processing of the 3D data set by the investigating radiologist is often required for confident diagnosis. Although image quality can almost always be improved by obtaining subvolume MIP images or by manuallyediting the entire data set, source images or thin-section reformatted images should also be examined routinely since small vessels, arterial dissections (Fig. 4) and non-occlusive thrombi may easily be missed on MIP images alone.

Fig. 4a, b. a Coronal MIP projection b Coronal source image from the 3D CE MRA acquisition reveals a right carotid artery dissection (arrow) not seen on the MIP projection
Carotid Artery Dissection

Fig. 5a-c.The MIP reconstruction (a) clearly demonstrates a carotid artery stenosis. However, quantification of stenosis is best performed on MPR images (b, c) perpendicular to the vessel axis (arrowin c, stenosis of internal carotid artery in the axial plane)

Mra Carotid Arteries
Fig. 6. Surface rendered view of MRA of internal carotid artery. The voxel size is 500 microns

Fig. 5a-c.The MIP reconstruction (a) clearly demonstrates a carotid artery stenosis. However, quantification of stenosis is best performed on MPR images (b, c) perpendicular to the vessel axis (arrowin c, stenosis of internal carotid artery in the axial plane)

The loss of image contrast on MIP reconstructions is most apparent at the edge of vessels. The signal intensity near the wall of a vessel is less than that at the center since pixels at the edge are partially volume averaged with the low intensity background. If the vessel is small (i.e. less than 4 times the voxel size) it will appear artificially narrowed. Vessel diameter is thus best measured on the individual partitions rather than on the projections (Fig. 5).

Was this article helpful?

0 0
Essentials of Human Physiology

Essentials of Human Physiology

This ebook provides an introductory explanation of the workings of the human body, with an effort to draw connections between the body systems and explain their interdependencies. A framework for the book is homeostasis and how the body maintains balance within each system. This is intended as a first introduction to physiology for a college-level course.

Get My Free Ebook


Post a comment