CHICAGO—In the debate over the efficacy of tomosynthesis or standard digital mammography, the winner is tomosynthesis, according to researchers who presented a study here at the Radiological Society of North America Annual Meeting.
The clinical bottom line: After one year of using tomosynthesis, fewer patient recalls were required and, at the same time, use of the newer technology resulted in detection of both more cancers and more dangerous cancers, said Emily Conant, MD, Chief of Breast Imaging at the Hospital of the University of Pennsylvania.
She and her colleagues performed digital breast tomosynthesis on all 15,632 mammography studies over a 17-month period. The results were then compared with recalls and cancer detection rates observed among the 10,752 digital mammography studies in the previous 12 months.
There were 1,117 recalls with digital mammography, a rate of 10.4 percent. There were 1,371 recalls with tomosynthesis, a recall rate of 8.8 percent, which translated to a reduction of 15.6 percent in recalls.
In the digital mammography group, 46 cancers were detected, including 32 invasive cancers—a cancer detection rate of 4.28 per 1,000. With tomosynthesis, 82 cancers were detected, including 59 invasive cancers, yielding a cancer detection rate of 5.25 per 1,000. That translated to an increase in cancer detection of 22.7 percent. Speaking at a news conference, Conant said the figure did not reach statistical significance due to low numbers of cancer detected in both groups. “We would have required 10 times as many patients to have achieved statistical significance for cancer detection,” she said.
The detection rate for invasive cancers was 2.88 per 1,000 with digital mammography and reached 3.77 per 1,000 studies with tomosynthesis—a relative increase of 30.9 percent. “These are the cancers we really want to detect because these are the killer cancers that go on to metastasize,” she said. “We very much wanted to detect more of these cancers.”
The positive predictive value was 4.12 for digital screening and 5.9 for tomosynthesis, a 45 percent increase in the key ratio of recalls versus cancer detection, and that was improved and was significant, she added.
“In our natural study for the overall group we observed a decrease in recall rate, an increase in cancer detection rate, and an increase in the positive predictive value of tomosynthesis,” she said.
The researchers also stratified patients by breast density and again found that there was a decreased recall rate for all but the most extremely dense breast category, and increased cancer detection as well.
“Digital breast tomosynthesis is a new, innovative technology for breast cancer screening, with a procedure very similar to that of a regular mammogram,” she said. “If, as a patient you were not told you were having this, you probably would not know it. It still requires breast compression and requires three to four seconds for each acquisition.
“The novel part is that multiple low-dose images are obtained across an arc of the breast. Those images are then reconstructed to get a three-dimensional rendition of the breast, and the radiologist can scroll through the imaging. Then with the same compression, a regular two-dimensional mammogram is performed. This is a combination of 2-D and 3-D, the way we are doing it in the United States.”
EMILY CONANT, MD
By comparing the images, radiologists can investigate if a suspicious image in the 2-D scan is a problem. Conant illustrated how one suspicious lesion was ruled out when the successive images acquired through tomosynthesis showed that the lesion was normal.
“We are now using tomosynthesis clinically, although we are not receiving any additional reimbursement for tomosynthesis. There is a cost for upgrading the equipment and software. The radiation dose to the patient is about twice that of a regular mammogram. The dose is still less than the maximum dose of radiation allowed for each breast. The cost to the patient is not increased at our site.”
She noted that early studies with digital breast tomosynthesis had suggested that use of the new technology can reduce the number of false-positive scans and that there were also early studies that indicated that the use of tomosynthesis might be able to increase the cancer detection rate.
And while different trials have shown decreases in recall rate and increases in cancer detection, she said that all those previous studies may have been subject to biases that resulted from sequential screening methods used in the studies. Her research team wanted to know what happens in real life.
She said that her results were similar to findings in a large Norwegian clinical trial of tomosynthesis. “It is important to reproduce findings to be sure that tomosynthesis is as good as we think it is.”
“We did all out screening with tomosynthesis. We started one morning with tomosynthesis and there was no other type of screening going on. Our population of women remained very stable between the one year prior with regular digital mammography and for the women screened through just tomosynthesis.
DEBRA S. COPIT, MD
“Breast density is a hot topic right now,” Conant continued, but when patients were stratified for breast density, recalls went down with tomosynthesis except for the category of very dense tissue. She noted that only about two percent of women presented with very dense tissue breast in either group.
Cancer detection and invasive cancer detection rates increased across all density categories—including very dense breast tissue; the positive predictive value also improved with tomosynthesis. “This tells us that tomosynthesis actually improves outcomes for all breast densities,” she said.
She said the researchers believe that tomosynthesis will be more rapidly accepted after the conclusion of clinical trials at multiple institutions that will show reproducibility of results across practice types and patient populations. Follow-up studies are necessary to determine if there are interval cancers that occur and to determine the full sensitively of tomosynthesis in finding cancers and preventing disease.
“What I like about tomosynthesis is that it is a very accessible platform, based on something that is economical and thus can be put in many places, whereas magnetic resonance imaging is costly and people are not going to be uniformly screened with MRI,” she concluded.
‘Ground-Breaking Work for This Modality’
Commenting on the results, Debra S. Copit, MD, Section Chief of Mammography and Director of the Gershon-Cohen Breast Clinic at Jefferson Medical College in Philadelphia, said, “I think this is ground-breaking work for this modality. We have been using tomosynthesis for about five years,” and think it should be considered standard of care. I can't imagine reading a mammogram without tomosynthesis.
“We do not charge patients extra for tomosynthesis at our facility because we believe in it,” she continued. “Studies like this one by Dr. Conant show that it is valuable for our patients. The radiation dose remains within what is considered acceptable. We are still relatively early with this technology, but this study is really significant in that it is a pure tomosynthesis population.
“The question of recalls is huge. We are most criticized for creating anxiety over false positives. If we were calling back fewer patients and not finding more cancers, that would be a problem. But every study, including this one, shows a decrease in call-backs and simultaneously an increase in cancer detection. I can't imagine anything better for the patient. It is not a panacea, but it sure improves on what is the gold standard right now,” she said.
Copit noted that there is an ongoing large multi-institutional clinical trial that may resolve some of the FDA and insurance-reimbursement questions.