Integrated Visualization of Physiologic Data in Cardiovascular Applications

Visualization of physiologic information is an important aspect of cardiovascular evaluations. Traditionally, physiologic information acquired by MR has been stored and displayed as separate image datasets. However, the benefits of both spatial and temporal correlation between physiologic and anatomic data are lost to the operator and the clinician. We have developed an integrated cardiovascular display that seamlessly presents physiologic data such as flow velocity, local blood oxygen saturation, and other data as a color-encoded overlay on anatomic images. The system uses a JAVA interface that allows, for instance, display of real-time color flow data sets at up to 20 frameisecond. Introduction Visualization of physiologic information is an important aspect of cardiovascular evaluations: The color overlay of flow information over anatomical data has been routinely used for some time in ultrasound. Several researchers have followed ultrasound's example and color mapped flow data in MR images [1,2]. However, most physiological information is still displayed separately from anatomical information, and when the information is combined using color overlays it has mostly been with applications tailored to a particular case. We have developed an integrated cardiovascular display that seamlessly presents physiologic data such as flow velocity, local blood oxygenation saturation, and other data as color-encoded overlays on anatomic images. Methods The display tool allows a clinician to view a color overlay on top of an MR image in real-time. The visualization tool runs as part of the iDrive real-time interface (GE Medical Systems, Milwaukee, WI). Most of our development was oriented towards visualization of color flow, but the tool can be utilized for the display of any relevant image combination. We will describe two particular use cases of the visualization tool: flow and blood oxygen saturation visualization. During a real-time flow study, typically the scanner operator will navigate using only magnitude images until the desired anatomy is located. Following that, a phase contrast acquisition is used to obtain both anatomical and flow information. When the color flow mapping is activated, the system waits until interleaved velocity encoded and a magnitude images have been acquired and overlays the color mapped velocity information on top of the magnitude image in real-time [l]. Controls are provided to set thresholds for both the magnitude and velocity values that should be color mapped. This allows the color to be mapped only in regions that contain physiologically relevant information. Currently, color overlays are not saved along with the images, but the visualization tool can extract the information and recreate the overlays from the images stored in the database. Non real-time images, such as gated phase contrast studies, can also be visualized while in this mode of operation. In non real-time mode, the tool will load a complete series and play it in a cine loop at a configurable frame rate. Once the images have been loaded and are being displayed, the same overlay tools are available as those in the real-time interactive version of the tool. For the flow visualization studies we typically use the color maps from ultrasound studies with the intention of presenting information to the clinicians in a familiar format. Several other colormaps are available, and the visualization tool was designed to facilitate the incorporation of additional colormaps and the customization of existing ones. As a second example, blood oxygen saturation (%02) can be calculated from T2 maps[3], acquired in a breath-hold following realtime localization. The resulting %02 map can be color mapped and overlaid onto one of the T2-weighted source images for anatomic reference. As with flow, the interface can be customized to adjust the mapping %02 value to color. To localize %02 information to vessels, T2 thresholds or flow thresholds from complementary data sets can be used. Results a n d Discussions The images show the value of color overlay on top of gray scale images in the clinical analysis of data. Figures 1 and 2 present axial images through the pulmonary artery (PA) in healthy subjects. In Figure 1, flow information is overlayed. For the gray scale presentation here, we map superior flow as white and inferior flow as dark. In figure 2 overlay intensity increases with %02. Since the visualization tool is not linked to a particular data type, it can be used to generate color overlays over any image that is acquired in real time. Its usefulness is not limited to flow visualization but will grow as the need for color mapping of realtime images grows. The visualization tool was designed to run as part of the standard iDrive interface. We tested the tool performance in a system with an Octane (Silicon Graphics, Mountain View, CA) 1270 MHz IP30 processor with 512 Mbytes ofmemory. With this configuration, we achieved rates of up to 21 framesis on a 5 12x5 12 display for the flow visualization. For single image overlay we achieved frame rates of up to 24 framesis. The rendering of the images onto the screen is consuming a considerable part of the processing time. By reducing the size of the display the 256x256 we expect to achieve higher frame rates. In the current implementation all real-time images are reconstructed to 256x256. Reconstructing to a smaller matrix size will also reduce Figure 1: Flow overlay in the pulmonary artery region Figure 2. %d)2 map oterla! in the pulmonarv ar leq region Conclusion The display of physiologic data over anatomic information is an important tool used successfully elsewhere in cardiovascular imaging. We developed a flexible display to specifically take advantage of the myriad of physiologic information available to MR. This real-time tool is able to handle static and dynamic studies either in real-time or from stored data. The achieved display frame rate of 20 U s is suitable for the display of real-time images even under conditions of stress. This tool can be extended to other clinical applications mapping physiological information such as strains, T2* maps for fMRI, as well as spectroscopic information. References: 1. Nayak, et al Mag Res Med 2000,43:251-258 2. Riederer ,et al Radiology 1991, 181:33-39 3. Wright et al, JMRI 1991, 1:275-283