Mammography CAD Comes of Age

As new CAD technologies emerge and hint of increasing clinical confidence and workflow advantages, the granddaddy of CAD applications—mammography CAD—continues to prove its merit in a shifting breast imaging world. The radiology community has accepted Digital Mammographic Interpretation Screening Trial (DMIST) results and looks forward to new imaging solutions like breast tomosynthesis. During the upheaval, one standard has remained steady; CAD remains a mammographer’s good friend.

Since its introduction in the analog era, CAD has provided a critical second eye, helping mammographers detect more cancers, earlier and in a cost-effective manner. It’s a win-win situation. As more sites transition to digital mammography, CAD transcends its early promise and delivers additional workflow benefits. This month, Health Imaging and IT visits with two mammography CAD pioneers to learn more about the benefits of CAD in the digital era.

Workflow under the microscope

“Digital mammography CAD has simplified life enormously,” reports James Ruiz, MD, a radiologist at Woman’s Hospital in Baton Rouge, La., and user of Hologic’s R2 ImageChecker CAD. The radiology department at Woman’s Hospital first deployed CAD in conjunction with analog mammography in 2003. Analog mammography CAD workflow required a fair amount of technologist time and input. For each study, a tech digitized and analyzed the film and returned it to its jacket and hung the study.

In contrast, digital mammography CAD eliminates the manual processes to streamline workflow and reduce tech manpower needs, says Ruiz. The CAD system automatically analyzes and marks suspicious areas on the digital images without disrupting tech workflow to further boost the workflow benefits associated with digital mammography. Similarly, CAD improves radiologists’ workflow in the digital environment. Ruiz explains, “The CAD marks simply pop up on the digital images. Our eyes never leave the image, so eye fatigue is reduced.” In an analog department, the CAD output is displayed on a separate monitor or on paper, forcing the reader to take his or her eyes from the film to review another image.

The Center for Breast Care at Boca Raton Community Hospital in Boca Raton, Fla., employs a slightly different efficiency model. The center is an early adopter of mammography CAD and deployed iCAD Inc.’s SecondLook Digital shortly after the technology earned FDA approval in 2002. Prior to implementing mammography, the center’s radiologists double-read all mammograms. Since implementing SecondLook, the department has phased out double-reading. “CAD makes us more efficient,” states Kathy Schilling, MD, medical director of imaging and intervention. Today, CAD serves as the second reader, allowing the site to accommodate a substantial increase in mammography volume without a correlative increase in FTE radiologists.

The practice reaps additional efficiency gains by digitizing analog prior studies with iCAD’s TotalLook Mammo Advantage. The system helped the center eliminate roller boards to minimize the stress of the hybrid mammography environment and accelerate the digital implementation.

A final, critical workflow consideration is the looming mammographer crisis, says Schilling. The reality of mammography is that fewer radiologists are entering the mammography field. CAD provides a critical aid as the crisis hits because it can reduce double-reading while increasing the detection rate.

The clinical rationale

Workflow is an important consideration in radiology; however, patient care is the top priority. The research on CAD is clear. “CAD has proven itself time and time again,” asserts Schilling, “It helps find more cancers. At this point, we would not read mammograms without CAD.” Research shows the use of CAD can detect up to 20 percent more cancers, and help detect nearly 75 percent of actionable missed cancers.

Woman’s Hospital is one of a handful of U.S. sites that employs a CAD/double-reader model. The hospital’s radiologists have double-read all mammograms since 1988. “Our experience has been consistent with the literature. Double-reading increased our detection rate by about 10 percent,” reports Ruiz. The math has changed a bit in the last five years. Today, the contribution of a second reader increases lesion detection by about 6 percent. “This tells us CAD is helping the first reader pick up more cancers, but not eliminating the contribution of the second reader.” With CAD algorithms on a continuous improvement path, the double-read model could come to a close. “There will be a breakeven point for our hospital as the algorithm improves,” predicts Ruiz.

A financial plus

With any healthcare technology, hospitals typically consider the bottom line, weighing the patient care benefits of a given product with its cost. Mammography CAD is a clear winner. “Reimbursement is sufficient to pay for the software, help improve the thin return on investment (ROI) in mammography and augment mammography reimbursement,” explains Ruiz. In fact, some CAD sites have seen ROI in five months.

Improving CAD

CAD is smart technology, and the systems continue to learn, resulting in improved algorithms and fewer false positives. Digital mammography CAD delivers a fairly hefty workflow boost by minimizing manual processes; however, other advances are in the works. The latest version of ImageChecker, for example, displays positive marks with information about what prompted the mark. “It tells us what it’s being intelligent about,” says Ruiz. And SecondLook will include features that provide additional clinical information about a lesion in future versions.

On target

Mammography CAD passes radiology’s triple test, offering workflow gains, improving clinical care and boosting the bottom line. “CAD,” says Schilling, “has become the standard of care. Not using it is like being in a car without a seatbelt.”


The Next Horizon: Breast Ultrasound CAD
Mammography is the first line of breast cancer detection, but it is far from perfect. Many cases require follow-up imaging, beginning with breast ultrasound and proceeding to breast MRI. Each modality is a candidate for CAD. In the case of ultrasound, it requires more time to discriminate a lesion from normal tissue, so CAD could play an important role.

Take for example the B-CAD application from Medipattern Corporation. It is designed to analyze breast ultrasound images. Catholic Medical Center in Manchester, N.H., a pioneer in mammography and breast MRI CAD, is one of a handful of sites around the country evaluating a beta version of B-CAD. “We set out to determine if the technology could improve patient care and pay for itself,” explains Michael Cloutier, manager of technical services. The answer is a definitive yes. Medicare and secondary payors have reimbursed the hospital for the procedure, and the hospital has discovered additional benefits associated with the technology.

“B-CAD enhances our ability to communicate with referring physicians. The software creates a standard report with a common look to ensure a quality report with all of the necessary elements regardless of the radiologist,” says Cloutier. He believes report quality can be a key differentiator in the radiology market as many clinicians cite report quality as a deciding factor in referral decisions. That is, clinicians tend to refer patients to practices that provide high-quality, appropriate reports.

The next version of the software, slated for release later this summer, also promises to streamline workflow by automatically identifying, segmenting and classifying the lesion to reduce technologist input. B-CAD automatically sends the file to PACS as a DICOM object for radiologist review. Reporting is standardized and streamlined via the B-CAD template.

“This is a technology that makes financial and workflow sense, and it [will] improve patient care,” sums Cloutier.