Like any high-technology system, CR requires regular care to ensure the quality of its images and the reliability of its mechanical operation.
At the pair of CR-equipped hospitals belonging to the University of California at San Francisco, the quality control program for that modality begins with a daily practice of erasing the photo-stimulable phosphor plate inside each image-capture cassette.
"Plates, we discovered, are quite sensitive to prolonged exposure to fluorescent lightingthey readily pick up signal from those light sources and that can have an adverse effect on image quality," says Katherine P. Andriole, PhD, associate professor in the Department of Radiology at UCSF and bioengineering at the University of California at Berkeley. "We once inadvertently allowed a plate to be exposed to fluorescent lights for about 12 continuous hours. Afterward, we noticed the images produced from it contained significant noise.
"To prevent a recurrence of that problem, we decided to henceforth make it a practice to manually erase every plate at the start of each day. This way we're assured that every plate is primed for use when our technologists begin to make their rounds, because otherwise you can't know until after you've shot and processed your first batch of images whether someone accidentally left a plate in a place where it could pick up noise-accumulating signal overnight."
Once a week, the plates are physically cleaned, Andriole adds.
"Dust is a problem for CR," she says. "If dust gets inside the machine and the cassettes, the particles can create image artifact. That's prevented by cleaning. However, CR plates can't be cleaned in the same manner as would be appropriate for the screens in an analog x-ray machine. The main thing is to avoid use of cleaners that are water-based. Frequently, though, with the CR plates, all that's necessary to get them clean is to sweep them with a dry cloth. Doing this weekly is usually sufficient, but we'll clean them more often than that if we notice artifact showing up on our images between regularly scheduled cleaning."
Every month, the plates are carefully inspected for signs of cracking. Those with cracks present are immediately replaced, says Andriole.
"The machine that translates the captured image from the plate will automatically erase the plate after the reading process is completedbut, over time, after several thousand read-erase cycles, the plates become susceptible to development of hairline cracks," she warns. "Cracks in the plates show up on images, but they may not look exactly like cracks. They can mimic the appearance of other things. For example, the first time a plate of ours began to crack, the outline of that crack was picked up on the image and looked exactly like a surgical staple. We weren't aware that cracking could occur, so we assumed it was indeed a surgical staple that had been left inside the patient. Further investigation showed us that was not the case."
Most CR maintenance tasks can be performed in-house. Some, though, are best carried out by the vendor as part of a service agreement. One of UCSF's CR vendors pays a service call to the enterprise on a quarterly basis to check system calibration and make other necessary adjustments.
"Whatever preventative maintenance chores you do on your own should be based on the manufacturer's recommendations," Andriole offers. "However, the frequency of those chores will depend on the volume of imaging being done. The more examinations, the more preventative maintenance that will be required because of the increased wear-and-tear that goes along with higher utilization."
Adhering to good imaging practice can also help ensure the quality of output. According to Andriole, a mistake some CR-utilizing radiology departments make is to deliberately underexpose images (as a way of reducing the radiation dose) and then using the computerized manipulation capabilities of the modality to bring the output up to diagnostic quality.
"If images are underexposed, you'll see mottle or noise over the image data," she says. "This can be a serious problem if, for example, you're doing chest imaging of premature babiesthe lung disease pulmonary interstitial emphysema shows up as mottle, so, if you're shooting with lower radiation dose, then you're not going to be able to tell whether the mottle on the image is the disease or merely low-dose-related noise."

Rich Smith is a contributing writer for Decisions in Imaging Economics.