Higher digital mammography recalls represent temporary problem

As imaging centers convert their workflow to capitalize on the benefits of digital mammography, some radiologists are concerned about the possibility of an increase in the number of patients recalled. Discrepancies between prior screen-film exams and current digital images lead many radiologists to act "better safe than sorry" in their diagnostic interpretation of full-field digital mammography images.

Radiologists may find themselves recalling patients with benign abnormalities that would previously have been excused if imaged on the familiar screen-film modality. The ambiguity that prompts these recalls is attributed to a learning curve, as breast imagers become acquainted with the mechanics of digital workflow. The recall rate for FFDM will likely decrease, however, once radiologists and their staff become familiar with digital images.

Clinical experience in the community-based practice at Elizabeth Wende Breast Care LLC demonstrates a decrease in recall rates in years following FFDM implementation. From 2003 to 2006, FFDM recall rates remained steady as technologists and radiologists began the transformation to digital imaging. During this time, the FFDM recall rates remained lower than those for analog film and were below the national recall average of 10%.

In 2006, digital recall rates peaked, coinciding with the installation of three FFDM systems from multiple vendors. Despite the increase in recalls for 2006, digital was on par with screen-film recall rates at 8%. The FFDM recall rate decreased to 7.7 % in 2007 and is expected to continue to decline in 2008.

Within six months following digital implementation, our radiologists had established a comfort level in interpreting FFDM images, benchmarked by a decrease in the number of patients recalled who were imaged with FFDM. For those radiologists struggling with a continual increase in their recall rate, transitional issues may still be present in the imaging workflow (Figure 1).

One issue that may hinder the digital progression is the comparison of prior screen-film images with current FFDM images. FFDM studies are displayed on 5-megapixel monitors, whereas prior screen-film exams are hung on adjacent alternators. The large amount of time radiologists spend reading mammograms dictates that serious consideration must be given to the position and setup of soft-copy workstations.

As time goes on, studies from outside facilities will likely still be sent as hard-copy images. Therefore, PACS stations need to be situated next to light boxes so radiologists can compare exams. Computer monitors should be strategically positioned to avoid excess light reflected onto the monitors from nearby viewboxes (Figure 2).

Depending on the radiologist's height and preference, monitors need to be raised or lowered to best match the individual's eye level. Ill-positioned monitors strain radiologists' eyes and inflict neck and shoulder pain, infringing on the reader's ability to interpret mammograms.

Ergonomically correct workstations have adjustable keyboard platforms installed to reduce wrist and elbow pain as a result of holding arms outstretched for prolonged periods of time. The top of the monitor screen should be at eye level, and the target viewing area should be 15º to 20° below eye level. The radiologist should be positioned an arm's length away from the monitor. Finally, an adjustable office chair will compensate for any remaining factors that may hinder the radiologist during interpretation.

In conjunction with adapting to reading digital images, radiologists have to become acquainted with the soft-copy workstations that display FFDM images. The PACS responsible for storing mammographic DICOM images and patient information is equipped with numerous tools to assist radiologists in diagnosis. Complete understanding of workstation capabilities is essential in order to fully exploit the benefits of digital mammography. Image manipulation post-acquisition reduces the need for additional views, as radiologists have the ability to magnify and pan unclear areas. Digital mammography is more tolerant of over- and underexposure of images, narrowing the margin of technical error and resulting in fewer recalls for repeat images.

PACS allows for more efficient communication between radiologist and technologist. Annotations such as circles marking suspicious areas and measurement tools are directly applied to the image. Radiologists can mark areas of concern and comment on what views they would like the technologists to perform without having to leave their office. Technologists can access this information at the image acquisition station inside the mammography room while the patient is present.

PACS allows radiologists to access all radiographic studies and pertinent patient information from linked databases. Images obtained from multiple modalities are accessible at one interface, enabling radiologists to compare mammograms, ultrasound scans, and MR images (Figure 3).¹ Prior screen-film images can be digitally scanned into PACS. These digitized images cannot be used for diagnostic interpretation but can be used when comparing prior studies with the current study (Figure 4).

Importing priors into PACS lessens the burden on radiologists making these comparisons, as it is much more comfortable for them to view all the images in one format rather than alternating their reading methods between light boxes and computer screens. Additional information such as physician notes, medical reports, and health history can be imported into PACS for radiologists' reference.

Installing PACS stations in each staff radiologist's office enables them to read images and obtain patient information from one location. Time previously wasted traveling to different areas within the Breast Center can be reallocated with a more efficient workflow.

The amount of new information and methodologies breast radiology practices need to learn as they proceed with their digital transformation means that issues will inevitably arise during the transition. Reading time will increase until radiologists become familiar with the visual appearance of the FFDM images and the tools of the soft-copy review workstation. Radiologists and PACS administrators will have to collaborate in designing hanging protocols to suit each radiologist's preference. PACS has the ability to display multiple hangings per patient, extending the reading time as radiologists include more priors in their diagnostic comparison.

Breast centers will see an initial increase in the number of patients they recall, but this will decrease once radiologists train their eyes for FFDM. A six-month adjustment period should be expected.

The 2005 American College of Radiology Imaging Network Digital Mammographic Imaging Screening Trial (DMIST) proved that digital mammography enhances a radiologist's sensitivity in detecting breast lesions in women under the age of 50 with dense breast tissue.² Digital imaging has the capability to radiographically penetrate dense breast tissue, resulting in the mammographic display of breast abnormalities previously unseen on screen-film images.³ A missed cancerous lesion was attributed to dense tissue in 26% of all breast cancers detected within 12 months after a negative test, according to DMIST.⁴ Breast imaging centers with mixed environments should grant women with dense breast tissue priority on digital imaging devices. Reviewing each patient's prior mammogram before annual imaging will allow technologists to triage patients to the modality best suited for their breast tissue.

Current literature highlighting the initial effect digital mammography has on recall rates should not hinder a radiologist's decision to implement digital technology into his or her practice. Abstaining from purchasing a digital unit solely out of fear of the learning curve can potentially result in undiagnosed cancer lesions.

A recent New York Times article, "In shift to digital, more repeat mammograms," included discussions of the disruption of workflow following FFDM installation, which may cause radiologists to be wary of the incorporation of digital technology. Another repercussion of the article was frightening the general public into believing that digital mammograms result in an increased number of recalls, increased use of extra views, and unnecessary biopsy procedures.

Radiologists must learn to compensate for their lack of familiarity with FFDM. If recall rates continue to grow, the workflow needs to be reevaluated. Consistently imaging patients on digital units creates a common ground for comparison with prior exams. This further attests to the repetition needed for radiologists to become confident in their diagnostic interpretation.

ACKNOWLEDGEMENT

Special thanks to Melissa Skolny, research coordinator at EWBC LLC, for all her help with this article.

References

1. Trambert M. Digital mammography integrated with PACS: real world issues, considerations, workflow solutions, and reading paradigms. Semin Breast Dis 2006;9:75-81.

2. Kerlikowske K. The mammogram that cried wolfe. NEJM 2007;356:297-300.

3. Fratt L. Optimizing workflow: creating the integrated digital mammography facility. Health Imag IT 2007;5:54-57.

4. Boyd NF, Guo K, Martin L. Mammography density and the risk and detection of breast cancer. NEJM 2007;356:227-300.

Dr. Destounis is a radiologist at Elizabeth Wende Breast Care LLC in Rochester, NY, and an associate professor

at the University of Rochester School of Medicine and Dentistry.