After a long haul, digital finally gains an edge

Digital mammography's superiority over film in cancer detection in select patients took industry watchers and researchers by surprise. While proponents hoped for a best-case scenario of a clinical draw between the two techniques, few expected digital to excel.

But that's exactly what occurred when investigators revealed the long-awaited results of the Digital Mammographic Imaging Screening Trial in September. The DMIST's favorable findings could spur digital's adoption and bolster research into related technologies. But its high cost may represent a stumbling block for cash-strapped breast imaging centers.

The DMIST found digital to be more accurate than film in detecting cancer in women under age 50, those who have dense breasts, and in pre- and perimenopausal women. The five-year study involved 49,500 women, 33 sites in the U.S. and Canada, and devices from four vendors.

In women under 50, the relative sensitivity of digital versus film was 49% and 35%, respectively, with digital offering a slightly higher positive predictive value. In pre- and perimenopausal women, sensitivities for digital and film were 47% and 38%, respectively, and the PPV for digital was again higher. In women with heterogeneously dense or extremely dense breasts, sensitivity for digital and film were 38% and 36%, respectively, with equivalent PPV.

"I think those of us who use digital knew it would prove to be better, but we didn't know it would be significantly better," said Dr. Emily Conant, chief of breast imaging at the University of Pennsylvania Medical Center, and a DMIST investigator.

Researchers have tried to quantify digital mammography's clinical benefit since its introduction as a prototype technology in 1996. The DMIST findings were unexpected in part because the major trials conducted to date, in the U.S. and Norway, failed to find a significant difference in cancer detection between the two techniques.

Edward Hendrick, Ph.D., a senior medical physicist in radiology at Northwestern University in Chicago and a coauthor of the DMIST study, said digital's overall accuracy in younger women and those with dense breasts wasn't just a small effect.

"It was a sizable effect," said Hendrick. "Having participated in prior studies that did not find a significant difference, these findings were a surprise."

Of the 49,500 women studied, about 65% fell into one of the three groups for which digital excelled, Conant said. The key common denominator is dense breasts, which can affect both younger women and those older than 50.

"That's a very high percentage, which says that digital mammography is valuable in detecting cancer in a large number of women undergoing screening," she said.

Digital has additional advantages over film. Digital images are easier to access, store, and transmit, and they can be acquired at a lower radiation dose without compromising quality. Moreover, digital mammograms read on soft-copy systems allow radiologists to view multiple processed versions of the same image and eliminate time needed for film processing, handling, and filing.

Whether the study results will encourage increased adoption of digital remains to be seen. Digital systems cost about $450,000, up to four times the cost of a film system. Many breast centers are strapped for cash, given the low reimbursement rates they receive for exams.

That hasn't deterred some sites from investing, however. An estimated 410 digital mammography units were shipped to U.S. customers in 2004, compared with about 200 in 2003.

"Cost remains a big factor," Conant said. "Individual sites will have to decide how much it is worth. But digital is the wave of the future, and this study is just the beginning."

About 8% of all mammography systems installed are digital, Hendrick said, noting that breast centers have some catching up to do if they are to serve all the women who could benefit most from the technology.

"Women who are knowledgeable and have dense breasts will probably ask for digital and may in fact change where they go for a mammogram if they can't get digital," Hendrick said. "It will be like in the early 1990s when women began asking if a site was ACR-accredited. They will ask if a site has digital and if they can get it."

But such demand could pose problems for breast imaging centers unable to justify the technology's high upfront cost.

Having established the clinical value of digital mammography, DMIST researchers are now finalizing other components of the study ranging from image quality to radiology interpretation of soft-copy exams.

The image quality portion of DMIST is an attempt to evaluate the types of tests necessary to ensure digital mammography exam quality. Not all tests required by the Mammography Quality Standards Act for screen-film mammography may be appropriate for digital.

A cost-effectiveness assessment is another component of DMIST. In addition, investigators are evaluating quality of life issues associated with patient callbacks based on false-positive results, and whether digital holds a socioeconomic advantage over screen-film.

DMIST's positive clinical findings could give related technologies, such as tomography, a boost.

"There are a lot of new things you can purportedly do with digital that may have an advantage but are unproved at this point," Hendrick said. "These need to be evaluated. Another aspect is looking more carefully at computer-aided detection, especially in women with dense breasts, and finding out how much CAD aids radiologists in screening."