As a piece of equipment, a film digitizer get little respect; but everybody seems to have one, and it does require proper care and feeding.

Radiographic film digitizers are such a low-profile piece of equipment that most digitizer manufacturers do not even sell them directly to end users. They are more commonly bundled as part of a picture archiving and communications system (PACS) vendor’s package or sold through the company that provides the software and hardware that the digitizer must have to transmit an electronic image. A digitizer without its front-end application is like a printer without a computer. It just sits there and nothing happens.

This is not to say that the end user, a hospital or an outpatient imaging center, has no control over what type of digitizer ends up on the floor. The two-level marketing strategy on the part of digitizer manufacturers may be breaking down as end users demand more control of their choices and unbundling, but so far it apparently is the way most digitizers are sold.

“If you go to a digitizer manufacturer and ask to buy a digitizer, they will say, Who are you and what do you want it for?'” says radiology systems expert Steve Munie. “Then you would go and have meetings with the company that is going to provide you with a workstation.”

Munie is PACS administrator for Wake Radiology, a 50-person practice in Raleigh, NC, that interprets for several smaller rural hospitals along with the major hospitals it covers. Munie buys film digitizers to be placed in the rural hospitals so that nighttime radiographs can be digitized for electronic transmission to the larger hospitals where Wake’s 24/7 radiologists will give them a preliminary interpretation. The actual films will be used for the primary read when the radiologist assigned to the rural hospital shows up for work the next day.

Bundling a digitizer to a front-end PACS or workstation not only benefits the digitizer maker, it may benefit the workstation vendor too

“Most of these vendors will recommend the digitizers that will offer them the best profit,” says Michael J. Cannavo, president of Image Management Consultants in Warm Springs, Fla. But he adds that digitizers “are such a miniscule part” of the overall radiology budget that nobody really pays much attention to how they are sold. Nonetheless, because digitizers do the important job of converting x-rays to soft copy, end users may be wise to keep the marketing process in mind when negotiating with PACS or workstation vendors about digitizing needs.

John Weiser, PhD, diagnostic imaging physicist at Xtria Healthcare, a PACS consulting firm in Frederick, Md, makes a further point. Because digitizers are coupled with a front-end application, the end user’s real choice is not just the digitizer. “Your choice is how well you like the software that goes with it.”


Cannavo says there are only five companies that make digitizers. All digitizers utilize one or the other of only two processes.

The first digitizers to come on the market used lasers to scan x-ray film line by line and translate light and dark areas on the film into numbers. Those numbers would be sent as electronic signals that would then be decoded into images on a PACS or workstation on the receiving end of the transmission.

Because the laser moves line by line over the film, the digitization is transmitted in great detail. Everyone interviewed for this story agrees that digitizers remain the “gold standard.” The trouble is they are expensive. Munie estimates they are “50% more expensive” than their newer competitors.

Laser digitizers are also more expensive to maintain. “The big thing is total cost of ownership over so many years,” says Cannavo. “Lasers have to be calibrated two to four times a year at a cost of $800 each time, so that’s an extra 10 grand over 5 years.”

There is an alternative that is nearly as accurate, less expensive, and easier to maintain. The latest digitizers use CCD technology, short for charge coupled devices. Cannavo says that for most purposes CCD images are indistinguishable from those produced by a laser digitizer.

“If you can’t see the difference, why pay for it?” he asks. “Technically, a laser has a wider optical density. With dark films you get a better resolution with a laser than with a CCD. But usually the very dark and the very light areas of a film are not those areas that make a difference in the interpretation. It used to be that laser digitizers were significantly faster, but CCDs have sped up and are comparable.”

Munie explains that CCDs do not scan film, they operate instead like digital photographic cameras, taking in the whole image at once. “When light hits the CCD element, it creates a voltage. A voltage is made out of the light and that gets turned into a number.” The entirety of those numbers is translated back into the digital image, which shows up on the monitor or workstation screen.

Munie says the first CCDs presented only 256 shades of gray. Current models present up to 4,096 shades, more than the 600 to 700 shades that the human eye can detect “at a brightness of 1,000 candelas per meter squared,” he adds. “But a CT or an MRI image is presented in 256 shades anyway so presumably there was never a loss. When you go to the plain film analog world, there are an infinite number of gray shades.” While more shades may marginally benefit interpretation in a very bright image, the added gray tones take longer to digitize, so there is a time tradeoff, Munie adds.

Cannavo says that while laser digitizers need fairly frequent maintenance, CCDs need close to no service. “There is a very high mean time between failures,” he says. “For some, it has been 35,000 hoursthat’s 4 years.”


Whether the choice is laser or CCD, end users of radiographic film digitizers need to pay attention to the front-end software that instructs the digitizer.

“The software does the calculation of the image,” Munie says. “If you just took the numbers coming from the digitizer directly, the image would be hard and grainy. There are algorithms in the software to make the image look better. Some companies may do a better job of this than others.”

Both the capability of the software and the features of the digitizer itself figure into the end result. “The software is becoming more important,” says Weiser. “There are certain features that I think are needed: The digitizer should be able to find the right window and level setting. If you are digitizing out of a film library where the films come from a lot of places, then the challenge is to find the lightest and the darkest parts of the image. The software has to analyze the film and find the best window and level setting so the initial display is not a washed out image.”

The digitizer should also be able to flip an image in case the film is inserted backwards. It should have a multiple film feeder so that film to be digitized can be stacked, Weiser says. The digitizer should also be able to recognize when there are multiple images on a film and be able to feed those images to the PACS separately, he adds.

“There also should be an ability to pull some kind of DICOM (Digital Imaging and Communications in Medicine) worklist back on to the digitizer so that whoever is operating the digitizer does not have to enter all the patient information [into the electronic transmission]. This manual entry is very prone to error. Sometimes, achieving this takes a special integration between the PACS and the RIS (radiology information system). The PACS doesn’t know about a study from 2 years ago, for example, before the PACS was installed. Sometimes you have to have a work-around back to that uncompleted state so that the digitized image shows up in the PACS as a completed study.”

Weiser says the end user needs to “get a prospective statement on the cost of ownership” for the digitizer in question.

Quality control is also necessary. “Dirt and dust can get in the mechanisms. We’ve seen things like paper clips get in the digitizer and cause problems, or somebody spills a soft drink. That is why you have to check the image on a routine basis,” Weiser says. “The ACR (American College of Radiology) recommends that some type of test pattern be run through it at least once a month.”

He cautions against using film that has been processed by a laser film printer. “The laser print is pixels and lines to start with and you can get some interference with the pixels and lines in the digital scan. The best test film is one made by photographic methods. We make our own. We x-ray our test pattern onto a piece of film and then put that film through the digitizer.”


What a digitizer is used for will determine not only what kind of unit is needed but also how many are needed. There is no rule of thumb that can equate imaging volumes to digitizer needs, says Cannavo, because no two institutions are likely to digitize with the same end in sight (see sidebar).

One thing that digitizers are rarely used for is to routinely convert film archives into soft-copy archives. The process is too time-consuming. At 1 minute or more per sheet of film, digitizing whole archives is not practicable.

“Back file conversion just costs way too much money,” Cannavo says. “That’s why it hasn’t gone anywhere.”

Digitizing priors is common, but the prior images on film are typically put through the digitizer the night before a patient is scheduled and only when the images are known to be needed, Cannavo says.

Digitizers are rarely used to create soft-copy images that will be used for primary interpretation. Digitized studies are almost always reference studies. Even at Wake Radiology, where they are used for nighttime teleradiology reads, the true films are given the primary read the next day.

“You could use the digitizer as a means of creating the soft-copy image [for primary interpretation],” says Wake’s Steve Munie, “but I would say if that’s your application, you should be doing CR (computed radiography) or DR (digital radiography). I would expect their quality would be greater than with digitized film. I would think no one would ever consider producing film and then digitizing that film just to read it. I couldn’t imagine that as an option, just go to CR or DR.”

Often one digitizer is all that is needed. Wake Radiology puts one CCD unit at each rural hospital from which night radiographs will be converted and transmitted electronically.

At Salinas Valley Memorial Healthcare System, which operates the 274-bed hospital of the same name in Salinas, Calif, there is also only one digitizer. It is a high-end laser digitizer, and it fills a big niche.

“Like a lot of radiology departments outside of the major medical centers, we were looking at having to go from the analog to the digital world,” says Salinas’ director of diagnostic imaging, Tom Burnsides. “We don’t have CR yet. We are straddling between film and CR over the next few months, and that sent us looking for a digitizer. I wanted something that was robust and that would not require a lot of hand feeding. It had to be something that wasn’t fragile.”

Salinas Valley uses its digitizer primarily to send film copies to referring doctors, and to digitize films coming into the hospital from outside sources so that they are part of a patient’s electronic medical record. Like a lot of sites, Salinas Valley is just developing its EMR. “A lot of the outside films we put on our network, and that way we have them all together,” Burnsides says.

To send out reference film studies, the image from the digitizer is fed to a laser printer, and that laser print is sent to the referrer, Burnsides says. He says this is a big improvement over the old analog film copier used previously. “That copier would take up 2 days a week of FTE. If the regular copy person was not there, someone else would have to go in and the image quality might be questionable. You want to put your best foot forward when sending out film, like a chest x-ray, a bone scan, a CT, or MRI. All those have different contrast parameters. You had to individualize each thing you copied. If you had a jacket with 100 films, that’s 1.5 minutes per film, plus the questionable quality.”

Salinas Valley settled on a laser digitizer. The digitizer itself was about $30,000 plus a 15% maintenance contract, Burnsides says. “It was not plug and play. It took at least 3 days to make sure the parameters with the printers were correct. We started being able to do only 14×17 but now we can do any size.” The vendor does a quarterly quality control review. “Paper clips sometimes were left on the film. The vendor had to come out and fix that,” Burnsides adds, “but we’ve never been down a day.”

Salinas put its digitizer inside its 3D laboratory, but only because it was convenient to have it there. Digitizers have a small enough footprint that they are easy to find space for. “I was going to put it in the darkroom, but it turned out the chemical fumes would have affected the electronics,” Burnsides says.

The hospital is also using the digitizer to put out soft copy for its NICU department as a pilot project to let those clinicians familiarize themselves with electronic reading on computers around the hospital.

“I wanted a digitizer heavy duty enough that if I had to copy a lot of film, I would be able to do that, but I don’t see copying film as a big crying need right now,” Burnsides says. “In fact, I look for the digitizer to take more and more of a background role when the CR unit comes in.”

That may be true for a lot of facilities. The niche for the digitizer may get smaller. But as long as there is film to be transferred to soft copy, there will always be a need for a digitizer. For the foreseeable future, every facility will have at least one.

George Wiley is a contributing writer for Decisions in Axis Imaging News.

A Different Use: Treatment Planning

In the Department of Radiation Oncology at the University of Pennsylvania, laser digitizers are put to a different use than at most facilities. The digitized studies are still used for reference, as they are in most applications, but they never go into a PACS. They are used for treatment planning and dosage calculation in fighting cancers.

“Those are the two aspects for which we use the digitizer,” says Indra J. Das, PhD, FIEPM, FAAPM, CSCI, professor and chief of clinical physics at the Philadelphia campus. “We take the digitized images and transfer them to a treatment plan. We do that when we only have a radiograph, which is about 20% of cases. The dose can also be calculated from the radiographic film when it is digitized.”

In these applications, it is not the digitizer that is important. It is the software that can translate a surgeon’s marks on the x-ray into similar calculations on the digitized soft-copy image. A different software package can calculate dosages based on what appears on the digitized x-ray, says Das.

“The laser digitizers are accurate but quite finicky. It can be a nightmare.”

Digitizers are used only rarely to produce soft-copy images from outside radiographs. “Our radiology is a filmless department,” says Das. “We were the first in the country. If we have outside film, we will make a copy. But we never use outside film to plan treatment.”