Radiographic film printers are necessary tools today, just as they were in the not-so-distant past when all images were captured on film. Even if a picture archiving and communications system (PACS) is in use at an institution, there are pockets of imaging demand that can be met only by printed film. Beyond that, printed film can play an important role as a fail-safe if electronic systems go down. Some disciplines, and indeed some institutions, still adhere entirely to printed film for diagnosis and treatment assessment, especially where image measurement is involved.
John Romlein, MS, is director of clinical projects for XTRIA Healthcare, a Maryland consulting firm that had its roots in a mega-PACS installation for the Department of Defense. The company still counts the DOD, with about 60 hospitals worldwide, as a major client.
Romlein calls today’s demands for printed film in an electronic imaging setting “film leaks,” and cites several examples involving sizes, shapes, and techniques, including the long images needed for scoliosis studies and the restrictions on soft copy mammography. “Mammography is an exception primarily because the spatial resolution of digital mammography has not been clinically proven and established. The FDA (Food and Drug Administration) approval was for soft copy in company with a laser printed image.” Scoliosis printers need to be able to produce three-foot images, which monitors cannot accommodate easily, Romlein says.
The PACS transition process itself leaves pockets of demand for printed film, Romlein says, especially in those cases where conversion of aging modalities, such as fluoroscopy technology, to digital imaging may be too expensive to undertake. The need for film printing, however, does not end with so-called leaks, according to Romlein’s colleague, John Weiser, PhD, a medical physicist at XTRIA. Weiser says when an institution has PACS and is “filmless,” there are still two big reasons why it needs printing capability. “One is backup for the PACS,” he says. “You may need to print directly from the modalities in an emergency. That could be for just a short period, but it could be a very high volume situation.” The second reason that printers are needed is that “you may need to print off the archive to someone who has no electronic access, such as a lawyer,” says Weiser. He says as a fail-safe, or what is commonly called a fail-over, the need for printers can be lessened somewhat by adjusting modality flow in an emergency. But, he adds, “You at least need printers hooked up to every modality that you have so you are not cut off.”
Demand for print film is common from orthopedists and surgeons who have to measure images, even though methods of electronic measurement have been developed and the dimensions drawn from printed film are not always reliable. XTRIA’s Romlein uses the example of a hip replacement. “The orthopedist has implant templates, and he is laying them on the film to determine the correct size. If the measurement is not accurate, then the surgeon will have to go into the operating room with two or three implants and actually test them in the patient’s socket. The implants can’t be used again once they have been opened, so the patient has to pay for all three. Laser printers are a lot like a TV set. They can have different sizes. The scaling can be completely off because the printer fits the images to its output format. We recommend using a sizing phantom, an object of known size that shows on the print.”
To assess printer needs, XTRIA’s Weiser says imaging volumes need to be determined for both backup and daily printing for outside clients. While nuclear medicine and PET scanners can require color capability, the whole list of the remaining modalitiesCT, MR, CR, DRcan be handled by the same sort of a printer, a gray-scale radiographic printer essentially printing in a 14×17 format, Weiser says.
Wet or Dry?Because they use chemicals to produce their radiographic prints, the so-called wet laser printers appear to have fallen out of favor. It is not that they do not do the job. They can be faster than dry laser printers. Yet, the majority of those interviewed for the accompanying article advise phasing them out and purchasing dry lasers instead. “One of the things we want to get away from is printers that use chemical processes,” says John Weiser, PhD, of XTRIA Healthcare, a Maryland consulting firm. “Chemical printers need to be used regularly or they have problems. They don’t do well if they aren’t used on a routine basis.” In a PACS scenario, where printers may be used mainly for emergency or outside physician demands, having chemicals dry up or deteriorate is not inviting. A second reason for favoring dry laser printers is chemical disposal. “I would recommend not buying wet lasers,” says J.A. Seibert, PhD, at the University of California at Davis. “They create an environmental issue. Disposal is expensive, and it’s controversial. They are much more expensive for upkeep.” -George Wiley |
CHOOSING A PRINTER
Once you have determined modality and volume demands, you can begin the process of printer selection. S. Jeff Shepard, MS, is senior medical physicist at the University of Texas MD Anderson Cancer Center in Houston. Because it is a cancer center, where use of prior studies and comparative tumor measurements are critical, Anderson continues to do all its diagnostic and treatment reading from hard copy film, Shepard says. This will change sometime next year when the center’s PACS system is completed and soft-copy review begins.
Because its demand for CT alone amounts to about 2,600 sheets per day, printing has been critical to Anderson’s efficiency and its level of care, says Shepard. About 4 years ago, the center found itself in something of a printer crisis. “We were having problems with downtime and throughput in our print operation, and it was killing us,” says Shepard. “If you can’t make film, you can’t read it. Whenever our existing laser printers would go down or produce substandard images, we suffered tremendously. I think we had 10 CT scanners operating through three different printers, and we were having problems with all three printers. We had reached the breaking point. We just couldn’t tolerate errors in print anymore.”
Because Anderson is part of a health care buying group, its choice for replacement printers was limited to three competitors. To decide between them, Shepard had each of the vendors install a machine side-by-side in a single room. “I had to beef up and rebalance the air conditioning to accommodate the three printers and two print servers,” he says. “Then I began to run objective performance tests for image quality, and for lots of things.”
Shepard says all three printers could be calibrated to produce images that were identical in terms of the specifications for density and contrast demanded by radiologists. But the printers had big differences by other measures. One required a special laptop to do calibrations for image quality and for DICOM (Digital Imaging and Communications in Medicine) settings to match modalities, while the other two had simpler calibration systems. Throughput speed was another criterion. “We just put a stopwatch to it and watched them come out, and that [speed] ranged between 50 and 130 sheets per hour. Those differences add up,” Shepard recalls.
Anderson also tested the machines for self-calibration. “You wanted film blackness very consistent for system output day in and day out,” says Shepard. “One machine measured itself on every film that came out and did a five-sheet average and then corrected itself if it saw a change in that average. Another machine measured a neutral density patch that it placed on the film, and the third measured the most absolute maximum blackness that it could find. The one using absolute blackness was hard to calibrate; it was all over the place. With the gray-scale measurement, it was much easier to control density. The other one measured 17 different shades of gray, but it did that only for the first sheet out of each box of film. But that recalibration turned out to be the most stable of all.”
In another test that turned out to be definitive, Shepard ran identical test patterns through the three printers. The printed test patterns were then measured. After that, they were left to hang on a view box for 6 weeks. “Guess what,” says Shepard, “they changed. One turned orange, one turned brown, and the one we ended up with had a slight color shift toward the blue. This was change in the base color of the transparencies. I had no confidence in the ability of these films to maintain their appearance over time. That was vital, because we pull more priors here than any other institution in the country except for the other big cancer clinics.”
In-House or Outsource Service?Radiographic film printers, like most machines, need to be watched and frequently adjusted. “Laser film tends to drift, and film batches are different,” says John Romlein of XTRIA Healthcare. “You need to calibrate the laser beam to the film’s sensitivity. To me, that’s an in-house job. I would think the physicist would come around periodically and make a test pattern and compare data and contrast and resolution, making sure there are no artifacts like scratches or marks from a crack in the delivery rollers.” But small and even not-so-small institutions may not have a medical physicist on staff, or even a trained technician who knows how to calibrate the printers. Of course, the printers do have self-calibration features. When those do not make the proper corrections, the institutions without trained in-house staff have to call in the printer vendors. But reliance on vendors alone can be costly, result in delays, and potentially could result in a lower level of health care. S. Jeff Shepard, MS, medical physicist at the MD Anderson Cancer Center in Houston, says reliance on vendors and lack of trained in-house staff lead to a complacency about printer quality control, and he says this is worrisome. “Most institutions never measure the quality and consistency of their image printers,” he says. “They mistakenly assume that this never changes. If the quality drifts so far that they can actually see a difference in their images, they call the vendor for service. However, since the quality changes slowly over time, subtle pathology that was apparent a year ago may no longer be visible, and vice versa. This could lead to a misdiagnosis.” Shepard adds, “Recently, the ACR (American College of Radiology) has added routine hard-copy image quality assessment to its list of items that must be addressed in order to obtain accreditation in its programs. This is a very positive step. Still, not all institutions are willing to pay the price to meet the ACR’s standards by establishing a comprehensive QC program. Unless it is mandated by law, as with the Mammography Quality Standards Act, or required by insurers, with Medicare usually leading the way, most institutions choose not to apply for ACR accreditation.” From his point of view, Shepard says, lax QC with printers is risky. “A well-designed QC program is worth the expense in the avoidance of unnecessary costs associated with the management of patients with misdiagnosed disease, not to mention legal issues,” he says. “Quality control should be aggressively supported as an essential tool to hold down health care costs while improving an institution’s standing.” But how can small institutions or those pinched for funds, as many are, afford to hire medical physicists or trained QC technicians? John Weiser, PhD, a medical physicist at XTRIA Healthcare, suggests sharing physicists or hiring them as consultants. “A vendor of a particular product could do QC for that particular product,” Weiser says. “The challenge is to do it for all the imaging output that you have. Getting everything matched up is usually the issue. My advice for a small institution would be to have a physicist come in and do a QC assessment of all their equipment. The physicist could work in conjunction with the service engineer and maybe come back to check everything on an annual basis.” -George Wiley |
Shepard also tested the machines for DICOM communicability with the various modalities, and for a range of other factors including formatting, images allowed per page, error handling, lookup table selection, and support for multiple modalities, a criterion for which all the printers were “very robust,” he says. Shepard also looked at fail-over redundancyor how well images from various modalities could be shifted between printers in an emergency.
Anderson now has 12 of its chosen printers, six located in three central print areas on its campus, handling hard copy production from 46 pieces of diagnostic equipment. Approximately 4,000 sheets of film are printed daily. There are 13 DICOM print servers connecting modalities to the printers. The system has been in place about 2.5 years, and administrators and physicians at the cancer center are “extremely happy,” Shepard says. The old printing nightmares that had plagued the institute are gone. This has been particularly noticeable in CT, which is the modality in heaviest use, Shepard says. “I have four printers and six print servers for my CT. If a server goes down, the system automatically switches to another one, and the same thing happens if a printer goes down. We have increased our throughput in CT by about 30%, which in our operation is worth about $10 million per year.”
On the cost side, Shepard says the $75,000 cost of each printer, while sizable, pales when compared with the cost of film. Film prices ranged from $1 to $1.35 per sheet, he says. When printers are purchased, film is purchased as well, he says. “We were able to negotiate a very good price.”
INTEGRATION
While printers need to be evaluated on the quality of printed images, they also need to work in concert with modalities and PACS installations. Printers can be connected directly to a modality, or the link with the modality can, and most often does, run through a DICOM print server that allows images from the modality to be routed to any of a number of printers. Images stored on soft-copy archives also have to be printed. It is essential that printers be integrated with these other systems.
“If you have a whole bunch of printers networked, the images from any one of the printers should look identical,” says XTRIA’s Romlein. “Any imaging device should be able to print off any printer, and that’s the whole idea with the DICOM gray-scale calibrations scheme. The images should look the same across all of the monitors and the printers.”
J.A. Seibert, PhD, is a medical physicist and a professor of radiology at the University of California at Davis, near Sacramento. Even though Davis has been completely digital in radiology for 3 years, Seibert has to contend with pockets of print need like everyone else. The demand at Davis for film is total in the OR and strong in orthopedics. “We probably will still print film for a long time,” Seibert says. “I think the market for dry lasers is actually going up.”
Seibert says working with the different vendors of modalities, printers, print servers, and PACS is essential because all of the devices must interface to work properly. “The key aspect is to make sure the printer manufacturer knows the types of modalities you want to connect to. You can connect directly, but via a DICOM print server is the best way. Although it is a little bit slower, it allows you a lot more flexibility.” The printer and the DICOM server have to be brought to the attention of the modality vendors also so that the variations in DICOM can be calibrated between printer, server, and modality. All of the vendors have to work in harmony. “You have to go in and touch each of the modalities to let them know that the printer is there, and the printer has to know the modalities. It’s a two-way street,” Seibert says. “That’s a really important issue.”
QUALITY CONTROL
A final, important issue with regard to printers is quality control (see sidebar). How often do the calibrations on printers need to be checked? Who needs to check them? “It’s really important to have a QC program headed up by a physicist or a super-tech who knows how to run the calibration strips,” Seibert says. “On our dry laser printers in our ambulatory care center, where we are part of a digital mammography trial, we run test strips every day, but for our conventional wet (chemical) laser printers, we run strips every week. A physicist checks every printer at least once a year. Checking daily on the wet lasers is overkill.” On that score, XTRIA’s Romlein points out that just reading off prints amounts to a daily background check by the radiologist “in the same way that the physician is supposed to do a general observation of the patient as well as a specific observation.”
When institutions are operating mainly in a PACS environment with soft-copy reading, anomalies in printing processes can develop unplanned. According to David Quinlan, RT, informatics PACS manager, Beth Israel Deaconess Medical Center, Boston, making prints for outside referrers and other outside needs was “getting shifted to the file room staff.” That put too heavy a burden on untrained people, he adds. “The file room goes in naively and says What I see is what I get.’ Their job for years has been to take prints out of one folder and put them in another. Now you’re asking them to recognize certain structures.” The hospital now has shifted the printing back to the CT and MR departments where technicians are more familiar with the criteria for the images that need to be printed.
As more and more radiology takes place using soft copy, the push to reduce hard copy will continue. But hard copy will not go completely away. For all the reasons noted by the sources for this story, hard copy cannot go away, and so radiographic film printers will remain a staple in radiology departments. Those who must purchase them and use them need to remember that printers are complex, sensitive machines that demand attention.
George Wiley is a contributing writer for Decisions in Axis Imaging News.