Home

Cover Story

Table of Contents

E-Newsletter

Article Archive

Editorial Calendar

Datebook

Writers' Guidelines

Orgs/Links

Opinion Polls

Reprints

Forum

For other articles and previous issues click here.

November 14, 2005

DMIST and the Switch to Digital
By Dan Harvey
Radiology Today
Vol. 7 No. 23 P. 40

The Digital Mammographic Imaging Screening Trial (DMIST) found that digital mammo outperformed film in three groups of women. The common thread may be dense breast tissue, which researchers plan to investigate further. The other major issue is how the more expensive digital technology will penetrate the mammography marketplace.

At long last, the wait was over. In September, the much-anticipated results of the Digital Mammographic Imaging Screening Trial (DMIST) were made public. Although the study produced no earth-moving tremors, it yielded important new information about breast cancer screening and detection.

The DMIST study, the first large-scale, multicenter study designed to compare the diagnostic accuracy of full-field digital mammography (FFDM) with film/screen mammography (FSM), involved nearly 50,000 women screened at 33 sites. Researchers found that the two technologies performed equally as well in the overall study population. However, FFDM performed statistically better in detecting cancer in three patient subsets: women younger than 50, women with dense breasts, and women who had not yet experienced menopause.

“We’re not really sure why it worked best for women in these groups, as that wasn’t one of those things we could test in the study, because we didn’t know what the results would be in the end,” reports lead investigator Etta Pisano, MD, FACR, professor of radiology and director of the Biomedical Research Imaging Center at the University of North Carolina School of Medicine in Chapel Hill. “But we are now looking more carefully at that, and we believe that breast density is probably the common denominator in those groups. It was the driver of the results.”

In women younger than 50, FFDM detected approximately 15% to 28% more cancers. In women with dense breasts, digital mammography picked up 15% more cancers—a significant finding, as women with dense breasts tend to be at higher risk for breast cancer. In premenopausal women, FFDM detected 21% more cancers.

According to the National Cancer Institute, which funded the study, women in the three subsets included approximately 65% of the women in the trial. As the researchers suggest, as FFDM finds more of the cancers causing death in the three subsets, the technology could save more lives.

Pisano says, for her, the results yielded only one surprise: There were no differences in the number of false positives between FFDM and FSM. “We expected that there would be fewer false positives,” she says. “But I’d rather have fewer false negatives, which is what we found, and the same number of false positives.” In other words, FFDM demonstrated higher sensitivity with no change in specificity, and Pisano was pleased with that outcome. “That is the best result we could have gotten,” she adds.

Research Goals
The main purpose of the trial, which was coordinated by the American College of Radiology Imaging Network, was to compare the diagnostic accuracy of digital mammography with FSM for the screening population of women without breast cancer symptoms.

Film mammography has proven effective in detecting early breast cancers and the breast cancer death rate has steadily dropped since 1990, but the clinician suspected that the technology has reduced sensitivity in women with dense breasts. FFDM had demonstrated potential in detecting more of the missed cancers, but DMIST investigators believed a large study was necessary to determine whether digital mammography was better than, or at least equivalent to, traditional mammography.

Launched in October 2001, the DMIST study involved 49,500 women and digital mammography systems from four vendors (SenoScan, from Fischer Imaging; Computed Radiography for Mammography, from Fujifilm Medical Systems; Senographe 2000D, from GE Healthcare; and the Selenia FFDM System and the Lorad/Trex FFDM System from Hologic). Screenings were performed at 33 sites in the United States and Canada.

Study Protocol
Women without symptoms or signs of breast cancer were enrolled at the time of their regular screening mammogram. Each participant underwent both digital and screen mammography. Both exams included two views of each breast. Two radiologists then independently viewed each woman’s images. The women had a second mammogram one year later.

Researchers were looking at diagnostic accuracy, sensitivity, specificity, positive and negative predictive value, and receiver operating characteristic performance. Specifically, each participant’s images were evaluated for the diagnostic accuracy of digital images compared with that of printed film, the effect of breast density on the diagnostic accuracy of digital mammography compared with film, and the diagnostic accuracy of both digital and standard mammography units.

In their paper published in the September 16 issue of the New England Journal of Medicine, the researchers reported that, for the entire study population, digital and film mammography exhibited similar diagnostic accuracy (difference between methods in the area under the ROC curve, 0.03; 95% confidence interval, -0.02 to 0.08; P = 0.18). However, FFDM accuracy was significantly higher than that of film mammography among women under the age of 50 (difference in the area under the curve, 0.15; 95% confidence interval, 0.05 to 0.25; P = 0.002), women with heterogeneously dense or extremely dense breasts on mammography (difference, 0.11; 95% confidence interval, 0.04 to 0.18; P = 0.003), and premenopausal or perimenopausal women (difference, 0.15; 95% confidence interval, 0.05 to 0.24; P = 0.002).

Researchers diagnosed 335 breast cancers during the 455 days after study entry in 2001. They found that, in general, cancers detected by either film or digital mammography were similar in stage and histology. However, lesions detected by FFDM and missed by film in women in the three aforementioned subsets included many invasive cancers and medium- and high-grade in situ lesions. Many of these cancers were confined to the breast at diagnosis. The researchers said these are the types of lesions that must be detected early if more lives were to be saved through screening.

Reaction
Pisano reports that, in general, reaction to the DMIST findings was positive, with many people saying the results were what they had expected. For those working at facilities already implementing FFDM, the findings offered no great surprise.

“I already felt that, through my own experience as well as anecdotal evidence, FFDM is the better procedure,” says Debra Mitchell, MD, director of Breast Imaging of Oklahoma in Edmond, a facility that has used GE digital mammography technology since its inception nearly three years ago. “The fact that FFDM works better in dense breasts in younger women tells me that it is superior. Those are the hardest of all of our mammographic studies to perform.”

Mitchell first used digital mammography at the University of Oklahoma, when the institution converted from traditional mammography. “We were looking for new equipment, and we could see that all the rest of radiology in general had gone digital, so it seemed the logical thing to do,” she recalls.

When she established Breast Imaging of Oklahoma with colleagues (Kelly McDonough, MD, Carol Sheldon, MD, and Ashley Magness, MD), the decision to implement digital technology was easy. “We had the option to go with digital or analog, and it was a unanimous decision for digital,” she recalls. “All of our technologists were in favor of it as well.”

The Cost Issue
Cost is viewed as a hindrance to utilization. An FFDM system can cost as high as $400,000 to $500,000—four times the price of an FSM unit. A facility would need to perform a high volume of studies to cover the expense, which raised the logical questions: If FFDM and FSM demonstrate equal effectiveness, what would be the advantage of replacing conventional technology with newer but more expensive equipment?

Pisano expresses a mammography community concern that women will demand digital mammography, pressuring facilities to purchase more expensive digital systems even though they may not be able to financially support them. “They’re worried that women will start clamoring for digital mammography, and there isn’t enough access to the technology yet,” she says.

To address this cost issue, DMIST researchers are conducting the adjunct “Cost Effective Analyses.” Results of the in-depth study are expected to be reported later this year or in early 2006, according to Pisano.

In the meantime, Mitchell indicates that—at her facility, at least—FFDM was the best diagnostic choice and proved itself cost effective as well. “Digital mammography is an expensive tool, but a digital machine can do the work of about two traditional mammography machines,” she says.

Vince Polkus, mammography applications product manager for GE Healthcare, echoes her observation. “Generally, all of our customers can replace two conventional systems with every one of the FFDM systems,” he says.

He indicates that there are inherent cost and productivity benefits with FFDM, especially for those who have the existing infrastructure to support digital technology. Costs, he says, involve more than the initial acquisition. Prospective buyers need to look at the overall life-cycle costs of a digital system. “Digital systems have a very long life compared to analog systems, and they have the ability to do more with less. You can reduce costs related to infrastructure, personnel, and archiving,” he says.

FFDM advantages over film mammography include easier access, storage, and transmission. In addition, the wait for film development is eliminated. In the long run, these can add up to savings, as they yield improvements in productivity and patient throughput.

Moreover, Mitchell believes the DMIST results will legitimize FFDM technology for insurers. “I think that, in some places, some third-party payors have held off paying for digital mammography pending the results of this trial,” she says.

Boosting Utilization
Mitchell believes those elements combined with the DMIST results could make FFDM more attractive to potential buyers and eventually increase utilization. “I think those that have been sitting on the fence, trying to decide whether this was a legitimate technology, will now cross over the line,” she says. Further, she believes study results will only help usher in the inevitable. “Eventually, people will have to migrate to digital, because everything else is going digital,” she says.

Likewise, Pisano believes the results will boost system sales. Before the study, she points out, there was little justification for purchasing digital over film—except perhaps for a “gut feeling” that FFDM was better. The DMIST results provide more solid validation. “Before, we had no studies that showed the benefit,” she says.

Polkus says GE definitely foresees a positive impact on sales as a result of the study. In fact, customer queries started immediately after the DMIST results were announced. “Our customers tell us that their patients are calling and asking if they offer FFDM,” he reports. “Because of all of the publicity, there’s heightened consumer awareness about the benefits.”

However, consumer awareness has raised a concern. Currently, only roughly 8% of all breast cancer screening centers in the country use FFDM systems. Healthcare professionals like Pisano worry that some women will forgo screening until digital mammography is more readily available. Pisano says it’s very important for women to realize that they need to continue screening, even if they don’t have access to digital technology. This especially applies to the women in the three subsets that would most benefit from FFDM. “Ideally, they would be able to get digital, but right now, that’s not realistic,” she says.

Evolving Technology
That will change as digital mammography evolves and as older film mammography units wear out and need to be replaced. “Essentially, screen film mammography has nowhere else to go,” says Polkus. “From a technological standpoint, you really can’t make the images any better. From a clinical and workflow perspective, you still have to handle film, which constrains the capabilities and productivity of a practice.”

But with digital technology, he says, so much can still be done to improve image quality and accommodate advanced applications. One potential advantage is that digital mammography could lead to usage of new imaging techniques such as 3-D imaging, dual-energy subtraction mammography, contrast-enhanced mammography, or low-dose mammography tomosynthesis. In addition, digital systems can feature add-on applications that enable better cancer detection, such as computer-aided detection (CAD) software. Polkus says that 80% of the full-field digital Senographe systems GE sells are equipped with CAD software. “As we look to the future, we see FFDM as only getting better,” he says.

Negotiating the Future
Results of the DMIST study have implications for patients, vendors, and clinicians. The researchers’ findings validated digital mammography as a viable diagnostic tool—one that is equal to and in some cases better than film-based mammography, particularly in cases involving women with dense breast tissue. As the entire world—and not just the radiology world—rushes toward digital, it appears inevitable that mammography will follow suit. The important challenge that now needs to be overcome is equipment costs.

“Clinicians need to work on ways to make the equipment purchases without breaking the bank,” says Pisano. “If we can’t do that, we are going to have a hard time replacing our equipment. So we are going to have to negotiate with the payors and the manufacturers. We’ll have to be good bargainers to afford the technology that women should have. This will not be an overnight transition. It is expensive and we will have to figure out how to pay for it.”

— Dan Harvey is a freelance writer based in Wilmington, Del., and regular contributor to Radiology Today.