|John C. Weiser, PhD|
The importance of maintaining image quality by careful quality control of the image acquisition device is well established. In a film-based environment, the end of the imaging chain was usually considered to be the film processor. If we could consistently produce a good quality film, our job in quality control was accomplished. Prior to the widespread use of digital imaging networks, the number of devices that had to be monitored in order to maintain image quality was limited and quite manageable. Enterprise-wide image networks improve the quality of medical care and efficiency of operation by providing greatly increased access to image information both inside and outside the medical facility. However, with increased accessibility comes increased complexity. Most of the PACS products currently on the market either are completely web-based or have a web-based component, allowing images to be displayed on any type of computer system. These systems can range from the low-end office desktop originally intended for e-mail and word processing to the dedicated diagnostic workstation in the radiology department. This diversity of display devices presents us with a new challenge for monitoring and maintaining image quality. Fortunately, recommended standards and commercial products are available to help us meet this challenge.
One of the first challenges in this integrated environment is how to maintain some consistency in the appearance of the image across a broad variety of computer displays, either a CRT or LCD. In general, there are four factors that affect the appearance of an image on a computer display. They are the black level, the white level, the relationship of the shades of gray between black and white, and the amount of room, or ambient, light that reflects from the surface of the display. The DICOM Committee has developed a standard, the Grayscale Standard Display Function (GSDF), which specifies a standard relationship for the shades of gray. Calibration systems are available that can measure the white level, the black level, and the shades of gray between; make an allowance for ambient light; and calculate a lookup table (LUT), which will convert that particular display in that particular room to a DICOM Standard display. As long as the monitor maintains its white level and black level, the display will maintain its GSDF calibration. However, displays do not normally maintain a consistent white level or black level over time. In the case of an LCD, the backlight will gradually lose efficiency, and the white level will slowly decrease. It also is not uncommon for someone to decide to take matters into their own hands and adjust the contrast and brightness controls of the display, completely destroy ng the calibration. So the display calibration must be verified at some predetermined interval, usually 3 to 6 months, and adjusted as necessary.
TIME PER MONITOR
|Screen capture illustrates the use of software to remotely monitor various workstation display features.|
Adjusting, calibrating, and documenting the results may take 30 minutes or longer per display, depending on the number of intermediate grayscale measurements that are made. This time per monitor is increased on a multi-monitor workstation by the need to match the white level and black level of all of the monitors. It becomes readily apparent that maintaining a GSDF calibration on a large number of displays in an enterprise PACS can become manpower intensive. Fortunately, features and software available on many of the newer medical-grade monitors make it easier for us to manage a comprehensive quality control program. The most significant new feature is a built-in photometer, or light meter, which monitors the white level of the display. This photometer can be either located in a corner of the front of the display or located inside the case, behind the LCD glass. This photometer then works in conjunction with special software that checks the white level at some predetermined interval, and adjusts the backlight setting to maintain the white level at a constant value. As long as the white level is maintained, the calibration is maintained. The marketing literature often refers to these monitors as “self-calibrating.” The software that comes with these advanced monitors usually includes additional features that help us maintain a quality control program and extend the life of the display. These features include the ability to automatically match the black levels and white levels of a multi-monitor workstation; the ability to automatically turn the backlight off at a certain time if it is not being used, and turn it on again prior to the start of normal working hours; and the ability to print a calibration report. Of course, with these advanced features comes increased cost of acquisition. The increased cost must be balanced against the decreased manpower required to maintain the display over its life cycle.
|One of the TG-18 test images used to implement a comprehensive image display quality program.|
The other new feature that has recently become available is remote monitor ing software. This software works with the workstation calibration software to provide feedback on the calibration status of a large number of workstations to a central location over the network. The software products are either a client/server application or a console product that uses the Simple Network Management Protocol (SNMP). These products are very useful for managing a large number of displays. For the high-end displays with the automatic calibration software, you can get real-time information on the operational status of the display. This information can include the current white level and black level, the date and time of the last calibration check, and whether the display is operating within specified tolerances. Other useful information can include the make, model, and serial number of the monitor; the software revision of the display driver; and the number of total hours the backlight has been in use. This information can be helpful in maintaining an accurate inventory and in predicting when replacement of the monitor should be considered. Other features to look for in this software include the ability to send an e-mail or a page if a monitor is failing, and the ability to produce consolidated quality control reports.
Assessment Tool Kit
A new report from The American Association of Physicists in Medicine (AAPM), “Assessment of Display Performance for Medical Imaging Systems,” provides a standard set of images and tests that enable assessment of display abnormalities. Instructions for obtaining AAPM Online Report #3 can be obtained from the AAPM web site, www.aapm.org.
These advanced features and the advanced software work well with the high-end monitors that we usually install in the radiology department and certain other critical locations, but they do not really help us with the majority of displays that are located on desktops, in treatment areas, and in clinics throughout the enterprise. It is important, early in the process of PACS implementation, to make a policy decision concerning the minimum hardware requirements for a workstation that is to be used for PACS viewing. While it is usually possible with most web applications to view PACS images on any workstation on the network, the institution should establish a policy for the workstations that are “officially sanctioned” for viewing medical images. These minimum requirements should include the minimum physical size of the monitor (ie, 17 to 19 inches diagonal), the minimum white level of the monitor for extended operation (ie, 170 cd/m2), and a display card that can accept a calibration lookup table if you intend to calibrate the monitor. Note that it is necessary to purchase a monitor whose published maximum white level exceeds your required white level, if you want it to be able to operate at the required level for any appreciable amount of time. You do not want to have to set the backlight to 100% at the outset to maintain the required level. You want to be able to achieve your required level with a reduced backlight setting, and then gradually increase the backlight setting over time to maintain the required level. The hospital policy can be flexible to allow PACS access on other workstations that do not meet the minimum requirements, but the user should be aware of the viewing limitations. Having specified some minimum requirements for hardware, it is still possible to end up with a formidable number of workstations that can be calibrated, should be calibrated.
While these workstations will require manpower-intensive manual calibration, there may be features of the remote-monitoring software that can be utilized on these lower cost monitors as well. While real-time feedback on the status of the monitor may not be possible, it may be possible to obtain a remote view of the date and results of the last manual calibration, and include this data in a consolidated QC report. If you have already decided to invest in the remote monitoring software for your high-end workstations, then you should check to see if it can also be used to assist in managing the calibration of these other workstations throughout the enterprise.
THAT’S NOT ALL, FOLKS
It would be nice if the GSDF was the only thing we needed to monitor to have an adequate display quality control program, but that is not the case. Proper GSDF calibration is just the beginning, not the end. We need to have a standard way to look for artifacts and display inconsistencies on our monitors. Over the course of time, the brightness of the display may degrade at a different rate in the center than at the corners, and present a significant variation in image appearance over the surface of the display. GSDF, by definition, is measured only in the center, so it does not give us any information on nonuniformity. Individual pixels may “burn out” with time. Manufacturing defects can cause delaminations, which appear as “smeared” pixels. CRT displays can lose their focus, causing a loss of spatial resolution. The American Association of Physicists in Medicine (AAPM) has recently released a report, “Assessment of Display Performance for Medical Imaging Systems,” which provides a standard set of images and tests that allow us to assess these types of display abnormalities. These tests are commonly referred to as TG-18 Tests, after the designated number of the AAPM Task Group that produced the report. Some of the calibration software applications now include the TG-18 test images and a method to systematically display and score them as part of the scheduled QC testing. These TG-18 tests provide the additional elements that are needed to implement a comprehensive QC program for medical image displays. Information on obtaining the AAPM report can be obtained from the AAPM web site, www.aapm.org .
So in the end, the mantra for PACS image viewing is this: “If you have a need to view an image, then you have a need to view it on a workstation that is subject to quality control.” The complexity and frequency of the quality control measurements will depend on the type of monitor, and to some extent on the intended use of the workstation. The requirements for a workstation that is used solely for viewing images along with a radiologist’s report may be less than those for a workstation on which clinical treatment decisions are made from the evaluation of the image alone. Our ability to manage a large-scale, comprehensive display quality control program can be enhanced by the purchase of monitors and software with advanced features for maintaining calibration and remotely monitoring status. Some of the newer remote monitoring applications also allow us to centrally obtain QC information on the lower-end monitors and produce consolidated reports. The final link in the QC chain is the TG-18 test, which gives us a standard method for assessing the overall display performance.
John C. 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.