Executives at Hologic (Bedford, Massachusetts) and iCAD (Nashua, New Hampshire) had plenty to say last week in response to a study on computer-aided detection (CAD), that study having the potential for doing considerable damage to a technology bannered as an important advance in the field of mammography.

The study authors conclude that rather than improving the accuracy of radiologists’ mammogram readings, CAD may instead make readings less accurate. The ability to improve on the detection of breast cancer has been the major claim for CAD by its developers.

The research was conducted at the University of California Davis Health System (Sacramento, California) and colleagues in the Breast Cancer Surveillance Consortium, which is sponsored by the National Cancer Institute of the National Institutes of Health.

The FDA approved CAD in 1998. CAD software analyzes the mammogram image and marks suspicious areas for radiologists to review, with the goal of assisting them to find serious tumors.

However, the results of the new study indicate that women who received mammogram screening at centers using CAD devices were more likely to be told their mammograms were abnormal: more false positives, requiring more callbacks for additional testing and having to undergo biopsy in order to rule out cancer. Findings appeared in the April 5 issue of the New England Journal of Medicine.

“First of all, what I would say is there’s a wealth of studies in the public domain that show that CAD finds more cancers and finds more cancers earlier, which is very, very important for women,” Ken Ferry, CEO of iCAD, told Diagnostics & Imaging Week, estimating the number of studies showing positive results from CAD in breast cancer at anywhere from 30-40. “So, it’s not about one study – it [should be] about many studies.”

That assertion was contradicted by Joshua Fenton, MD, of UC Davis Health System, and lead researcher in the study.

“Within three years of FDA approval, 10% of the mammography facilities in the country were using CAD,” he said, adding: “There had been no large-scale community-based review of CAD efficacy despite the rapid adoption of this technology, so we did this study to see if CAD was proving to be beneficial.”

Investigators looked at the use of screening mammography in 222,135 women who had 429,345 mammograms. They were observed from 1998 through 2002 and took place at 43 facilities in Colorado, New Hampshire and Washington states. The study included 2,351 women who received a diagnosis of breast cancer within one year after screening and also received a mammogram that did or did not use CAD.

Seven facilities, or 16% of the study sites, implemented CAD during the study period. With the use of CAD, 32% more women were recalled for more tests and 20% more women had a breast biopsy.

And according to the researchers, the study suggests that software had no clear impact on the early detection of breast cancer. It also suggested that the software may promote the detection of the least dangerous breast cancers, such as localized, in situ breast cancers.

In response, Ferry criticized the time frame of the study as a major weakness of the study, since the “most recent technology would be five years old, and the oldest technology would be nine years old. In a business that is about technical innovation in software, that’s a lifetime and more.”

He noted that iCAD, for example, is now on its seventh version of its software algorithm, having been introduced in 2002.

Ferry also noted that the CAD software was being used on analogue film-based mammography equipment, and while he said there’s “nothing wrong” with that, “the adoption of digital mammography equipment is white hot, because of very large studies such as DMIST that showed the superiority of digital technology in certain categories of women.”

Glenn Muir, executive VP and CFO of Hologic, joined Ferry in finding problems with the study.

“Clearly, we’re disappointed by the conclusions,” he told D&IW, saying that these conclusions are “really contrary to what I think radiologists are finding in their clinical practice . . . .”

Muir said that about one-third of screening mammography sites in the U.S. use some form of CAD.

“And if a radiologist has used it, chances are, if they’ve used it long enough, they’ve been turned on to an area that they may have overlooked without the second read,” Muir said.

Rachel Brem, MD, director of breast imaging and intervention at George Washington University Medical Center (Washington, DC), and a board member of iCAD, said this UC Davis study would in no way affect the use of CAD in her practice.

The study “definitely shows an improvement in cancer detection and it approaches statistical significance, even though the study was designed such that it’s stacked totally against CAD,” Brem told D&IW.

She argued that the radiologists who are interpreting CAD obviously have years of experience in reading mammograms “before CAD implementation, so the incremental improvement changes.”

“[This study] will not change my practice at all, and I think it would be a terrible disservice to women if this one study resulted in a decrease in the use of CAD,” Brem said. “It is an important, effective technology for the earlier detection of earlier breast cancer.”

However, Ferris Hall, a radiologist at the Beth Israel Deaconess Medical Center (Boston) told the Wall Street Journal that the study “is a substantial hit for this technology. This study suggests that [CAD] may actually be harmful.”

One problem indicated for the technology might be put under the category information “overload.”
According to the researchers, every time the CAD software marks a real cancer, a radiologist has to consider about 2,000 additional false-positive marks, making it “very difficult to distinguish between real cancers and benign tumors.” The authors estimate that for every additional woman diagnosed with breast cancer with the use of CAD, 156 women are falsely recalled for more tests and 14 had unnecessary biopsies to exclude cancer, biopsies that may result in new complications.

“It’s unfortunate that the use of the software has proliferated so widely before we are certain of its benefits,” said Fenton. “We need studies to determine if the benefits of the software outweigh its harms and costs. There is also the potential for new studies to improve the performance of CAD software.”

Ferry said that iCAD had not received any calls from customers on Thursday following major main-stream publication of the repot. And Muir said Hologic also had not received customer calls.

But both acknowledged being overwhelmed by investor enquiries.

Ferry said that iCAD’s customers providing mammography “take a very serious look” at the NEJM study but that the article was just one “data point” that these customers will put in the context of other studies available to them. “They tend to form their own opinions,” Ferry said. “And then I would say that they talk to their colleagues using the products locally.”