|
A 3D-image of the abdomen using a VitreaTM Workstation.
[Critical areas of eloquent cortex] Critical areas of eloquent cortex (blue region) that can be avoided in neurosurgical procedures, using the Stealth Station.
[PET and MRI fused image] PET and MRI fused image with intensity correlation using the Automated Image Registration (AIR) program provided by Roger Woods.
[PET and MRI fused image] PET and MRI fused image with intensity correlation using the Automated Image Registration (AIR) program provided by Roger Woods |
|
Medical Image Fusion
Technological advances in medical imaging in the past two decades have enabled radiologists to create images of the human body and its internal structures with unprecedented resolution and realism. State-of-the-art CT,* MRI,** PET,*** and other imaging devices can quickly acquire 3D images, and these images can further be computed to merge into a single volume thus combining the information of all modalities. The available computing power and sophisticated display software allow for the fused images to be captured on screen with both scientific authenticity and esthetic worth. Thus, recent advances in image acquisition, image fusion, and 3D-image display have radically changed our ability to visualize the inner workings of the human body.
Typically, this algorithm is used for pair-wise matches of different head images. Intensity matching uses mutual intensity information to co-register different images. The matched intensities may come from the same scanner (two different MRI scans acquired on different days, for example) or from different modalities such as MRI and PET. Similar image alignment methods can also be used to superimpose images with other information such as radiation dose, allowing radiation oncologists to see what regions will be affected by radiotherapy. Figure 2 shows segmentation of a brain image for radiation treatment planning Left: an example of a three-dimensional dose distribution calculated and overlaid on a segmented data. The automatically segmented structures are the brain stem (red), caudate (cyan), corpus callosum (magenta), putamen (blue), and internal capsule (green). The radiation isodoses shown are 70% (red), 35% (green), and 14% (blue) of maximum dose. Right: the three-dimensional dose distribution can be condensed into a two-dimensional graph, known as a dose-volume histogram. This graph is useful for dose-volume analysis for a given treatment, and may be used as input for calculating complication probabilities. Figure 3. Several software packages now enable users to manipulate and measure multidimensional, multi-modality image data, and analyze structure-to-function relationships. Different programs can be used for quantitative analyses, so that one may derive intensity measurements from various scanners. For example, intensity-based tissue classification and anatomical regions of interest can be defined with MRI, while metabolic activity from the same region can be measured with PET. Figure 3 shows PET and MRI fused image with intensity correlation using the Automated Image Registration (AIR) program provided by Roger Woods. Figure 4. Images are currently being used for image-guided neurosurgery by University of Iowa Health Care neurosurgeon Timothy Ryken, MD. Three-dimensional images from MRI have been used to visualize the cortical surface of the brain. This information can be used to avoid critical areas of the brain involved in movement or language. Figure 4 shows critical areas of eloquent cortex (blue region) that can be avoided in neurosurgical procedures, using the Stealth Station. Figure 5. Additional information from functional images, such as PET or functional MRI, may indicate which areas are important for vital functions, such as motor or language areas. Figure 5 shows PET image fused onto MRI image and presented in 3D using AnalyzeTM Software. Figure 6. Image-guided surgical navigation is becoming commonplace and "virtual surgery" can be done to rehearse surgery or plan surgical approaches. In the operating room, surgical probes, with sensors localized within the 3D space of patient's head, provide input to a computer-generated 3D-image derived from CT, MRI, or PET. Figure 6 shows a lead pellet in the orbit of the eye, localized using the Stealth Station. Contact information A clinical service for medical image processing has been established in the Laboratory for Imaging Applications in the Department of Radiology at the UI College of Medicine. An Image Processing Technologist performs the co-registration (fusion) of images from CT, MRI, and PET thus providing higher diagnostic value and improved treatment outcomes for a variety of medical conditions. Additional services of the laboratory include 3D-image rendering, multi-spectral image analysis, image segmentation (outlining of structures), assistance in surgical planning, and quantitative measurement of images. Physicians may put their requests through by calling the Laboratory at 319-384-8095. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 The images published in this article have been made possible thanks to Michael Vannier, MD, Sanford Meeks, PhD, Mark Madsen, PhD, Timothy Ryken, MD, Kurt Smith, DSc (Sofamor Danek, Inc.), Richard Robb, MD (Mayo Biomedical Imaging Resource, Rochester, MN for surface matching algorithm in Analyze(TM)), Roger Woods, MD (Automated Image Registration (AIR) UCLA for voxel matching algorithm), and NIH grant NS35368.
* Computer Tomography ** Magnetic Resonance Imaging *** Positron Emission Tomography
|
This site is optimized for IE 4, Netscape 4 or later with a window size of 800 by 600 pixels. Copyright © 1999-2001. The University of Iowa. All rights reserved.