One of the processes that changes significantly with a picture archiving and communications system (PACS) implementation is the method of delivering images and image quality outside the walls of the radiology department. Back in the days of film, controlling image quality was relatively simple. There was a limited image distribution system with centralized quality control. The focal point was the film processor. There were a limited number of these devices, and most of them were in the radiology department, convenient to maintain. When the film came out of the processor, that was essentially the end of the image production and display process. The quality of the image display was essentially fixed at the time the film dropped out of the processor, and it was as good or as bad as it was going to get for any future viewing. The image that was seen in the radiology department at the time of acquisition was essentially the same image that would be seen in a clinic a day later or a year later. Granted, the panels that were used to back-light the film for viewing had some effect on image quality, but the film itself was the overriding factor.
With PACS, our limited image distribution system becomes essentially limitless and our centralized quality control becomes very much decentralized. We no longer have the film processors conveniently located in the radiology department. Instead, we have possibly hundreds of computer workstations, each with a graphics card and some type of CRT or LCD monitor or monitors. These monitors are at the end of the imaging chain, and they determine the quality of the image every time it is viewed. Most PACS today are web-based or have a web server option that enables images to be displayed on computers with consumer grade graphics cards and monitors, not necessarily designed for optimal display of medical images. So our challenge as we upgrade or replace these desktop workstations is how to provide adequate image quality without breaking the bank. Fortunately for us, consumers like to play video games and watch DVD movies on their computers, and this consumer demand has driven the industry in a direction that provides us with some cost-effective options for clinical viewing of medical images. In order to make an informed purchase decision, we need some criteria by which to judge what is the best solution.
CONSIDER 5 CRITERIA
No. 1: CRT or LCD? The first option to consider is CRT or LCD. This decision is simple. Just walk into any computer store and look at the CRT offerings. I suspect that your choices will be few. LCD monitors provide us with a larger viewing area in a smaller footprint, are brighter, and provide additional advantages for long-term ownership, so we will limit this discussion to LCD monitors.
No. 2: How Big? Remember that our old film view box could display a chest film that was 17 inches high and 14 inches wide. The available viewing area of a computer display is usually reported in terms of its diagonal measurement, so a 17-inch LCD does not provide us with a 17-inch image. Consumer LCD monitors are designed to be used in landscape mode, meaning their long dimension is horizontal, so a 17-inch diagonal monitor will display a chest image that can be no more than 10 inches high, and will probably be smaller due to the area taken up by pull-down menus and icons that are part of the display application. There are some ways, which will be discussed later, that allow turning a monitor 90 degrees and using it in portrait mode. If the long dimension of a 17-inch diagonal monitor is vertical, then the chest image could be somewhere close to 13 inches high. For a somewhat higher investment for the monitor and graphics card, one can purchase one of the 21-inch diagonal LCD monitors that have recently become available to the consumer market. With one of these monitors oriented in portrait mode, the vertical dimension of the available display would be around 16 inches, so the chest image would be only slightly minified, compared to a film on the view box.
No. 3: How Many Pixels? The number of pixels in the display matrix goes hand in hand with the physical size of the display, up to a point. A 15-inch display will have a 1024 x 768 pixel matrix. A 17- or 18-inch display will be 1280 x 1024, and the consumer displays that are 19 inches diagonal and larger will most likely be 1600 x 1200. These matrices are sometimes referred to as XGA, EXGA, or UXGA, respectively. Another method of characterization, which has fallen out of common usage, is to list the number of thousands of pixels, using K as the abbreviation for thousands, in the longest dimension. Thus, 1024 x 768 is a 1K display, and 2048 x 1536 would be a 2K display. The more common method, driven somewhat by the way digital cameras are advertised, is to multiply the two pixel dimensions together, and then report how many million pixels, or “megapixels (MP),” are contained in the display area. Using this method, the 1600 x 1200 monitor has 1,920,000 pixels, and is described as a 2MP display. Diagnostic quality monitors used for interpretation of studies in radiology are also physically in the 21-inch diagonal range, but typically have a pixel matrix of 1536 x 2048 (3MP) or 2048 x 2560 (5MP). The pixel matrix determines how much of the image can be displayed on the monitor at full resolution. Looking at it another way, the pixel matrix determines how much downsampling of the original image has to occur if you want a static view of the entire image. Since a typical digital chest image contains about 5 million pixels, the 5MP monitors can display the entire chest image without any downsampling. On monitors with less than 5MP, we need to zoom in to a particular area of the image if we want to see it at full resolution, and this is a commonly accepted practice for both diagnostic and clinical use.
|John Weiser, PhD
No. 4: How Bright? The brightness setting of the monitor has an important effect on overall image display quality. The older unit of measure for brightness was foot-lamberts. The newer unit, which is used on most LCD monitor specifications, is candela-per-square-meter (cd/m 2 ). The American College of Radiology recommends that a monitor used for primary interpretation should have a brightness of at least 50 foot-lamberts, and this is equivalent to about 170 cd/m 2 . It is not unreasonable to also apply this standard to monitors used for clinical review outside of radiology. Important points to keep in mind when reviewing brightness specifications:
- The brightness specifications for consumer grade monitors are often “typical” values, meaning that some of the monitors may not be able to obtain the specified number, and others may be able to exceed it.
- The maximum obtainable brightness will decrease with time, as the backlight ages. Therefore, it is best to purchase a monitor whose maximum brightness is greater than the brightness at which you wish to operate it. In this way, you can operate the monitor with a decreased backlight setting when new, and then increase the backlight setting to maintain the desired brightness over time.
No. 5. What About the Display Card? Medical grade monitors are usually sold as a package, with a graphics card that has been tested to work properly with the monitor. This is not the case in the consumer market. Some points to consider when selecting a display card:
- Most consumer graphics cards do not directly support portrait mode, and instead rely on software translation to allow you to switch between landscape and portrait mode display. In some cases, using software translation can have an adverse effect on display performance.
- Not all graphics cards support monitor calibration. If you wish to use calibration software to standardize the display appearance, you need to make sure that the graphics card can be addressed by the calibration software.
In general, bigger is better, and brighter is better. You do not necessarily need to have a 21-inch monitor, but you should be aware of the amount of minification that will result from the use of a particular size of monitor. The 14 x 17-inch chest image was used as an example because it is the extreme case. If someone primarily looks at CT, MRI, ultrasound, or nuclear medicine studies, then the size of the monitor is not as critical. Brightness is equally important. A dark monitor yields a poor quality image, no matter how big it may be or how many pixels it can display. Consumer grade displays can provide adequate image quality for viewing PACS studies, but the selection must be carefully considered, in order to ensure that the final product meets the needs of the enterprise.
John Weiser, PhD, is a medical physicist and chief scientist at Xtria Healthcare, Digital Imaging Solutions, Frederick, Md, a professional consulting and managed services firm for digital imaging environments.