As the number and type of breast imaging modalities have multiplied and research has demonstrated their effectiveness, it may be time for women's imaging practices to consider the role of new technologies in breast cancer screening and detection.
In the last 18 months, regulatory, technical and legislative developments have occurred at a head-spinning pace. The FDA approved tomosynthesis, and an FDA advisory panel recommended approval of automated ultrasound as an adjunct to mammography for women with dense breasts.
From 2D to 3D
When the FDA approved breast tomosynthesis in February 2011, Houston Breast Imaging in Texas was poised to take advantage of the 3D development. Stephen L. Rose, MD, president and CEO of the practice, served as principal investigator in clinical trials of the technology. The practice has deployed four tomosynthesis systems, and its three affiliated hospitals own another four.
Many practices in the U.S. that have invested in tomosynthesis operate in a hybrid environment, with a single tomosynthesis system mixed with digital mammography, and have not devised a true tomosynthesis screening program. In contrast, by May 2011, Houston Breast Imaging began delivering tomosynthesis to its screening population.
In preparation for a clinical trial, Rose and colleagues have off-loaded 2D mammography data from 10,000 screening patients to a separate workstation to inform an "apples to apples comparison" of 3D and 2D exams. As the researchers evaluate the data, the early findings are compelling.
In the 2D environment, the 14-radiologist practice had a callback rate of approximately 10 percent. Since transitioning to screening tomosynthesis, its rate has hovered in the 5 to 6 percent range. Similarly, a study conducted at Yale University School of Medicine in New Haven, Conn., and presented at the annual meeting of the American Roentgen Ray Society in May, showed that the patient recall rate dropped from 11.1 percent with 2D mammography to 6.6 percent with 2D mammography and tomosynthesis.
"It isn't just that we are reducing the callback rate. When we call patients back, we don't have to do as many exams," says Rose. The conventional diagnostic model typically requires additional mammography views followed by ultrasound in most cases. With tomosynthesis, the patient typically proceeds directly to ultrasound and a same-day biopsy if needed.
Furthermore, positive predictive value has improved. "Traditionally, 20 percent of biopsies in a screening population are actually malignant. Our data show this rate is 45 to 50 percent with tomosynthesis," says Rose.
However, it is critical that imaging modalities detect malignancies of consequence to the patient's health, Rose says. "We are finding the same spectrum of tumors we see in the regular population, such as infiltrating lobular cancers or high-grade tumors. They are just smaller."
Rose and his colleagues had expected tomosynthesis to assist in the evaluation of women with dense breasts. A somewhat unexpected finding is its role in detection of cancers in women who don't have dense breasts. In some of these women, findings that seem to be an area of asymmetry have been diagnosed as 2 mm to 4 mm tumors.
However, Rose acknowledges the technology is not perfect. For starters, false-negative results are always a possibility.
Workflow presents another challenge. Tomosynthesis allows the radiologist to view data in 1 mm slices. In settings where radiologists can use the vendor workstation, Rose estimates that it takes 30 percent longer to review tomosynthesis images than conventional mammography results. In one location where the breast imaging workstation cannot display tomosynthesis images, data must be sent to a separate workstation. In that model, it takes two to three times as long to review tomosynthesis data. The situation is expected to improve when the tomosynthesis vendor releases a DICOM format for its data.
The efficiency hit is exacerbated by a reimbursement catch-22. The FDA requires that women have a standard mammogram in addition to a tomosynthesis study, yet the Centers for Medicare & Medicaid Services (CMS) and private payors do not reimburse radiologists to read tomosynthesis images. However, the increased diagnostic confidence in both negative findings and callbacks outweigh the downsides, says Rose.
MR, US & the intermediate-risk conundrum
Supplementing screening mammography with ultrasound or MRI improved breast cancer detection among