|
Introduction to digital mammography: The technologist’s perspective Kathleen Willison, RT(R)(M), Judith LaBella, RT(R)(M); Margarita L. Zuley, MD
Ms. Willison is a Consultant, Ms. LaBella is the Lead Digital Technologist, and Dr. Zuley is a Staff Radiologist at The Elizabeth Wende Breast Clinic, Rochester, NY. Ms. Willison is also the Director of Clinical Affairs, Koning Corporation, Rochester, NY, and Dr. Zuley is the Chair of the IHE Mammography Working Group. Last year, approximately 150 certified U.S. mammography facilities closed their doors. In April 2005, there were 9011 such centers in operation; by April 2006, that number had decreased to 8860.1 At the same time, however, the number of full-field digital mammography (FFDM) systems in use increased from 819 to 1331, and the number of facilities with at least one digital system jumped from 607 to 924.1 Whether the decrease in the number of mammography facilities is at crisis level or not, digital technology has the potential to increase access and exploit existing resources. Benefits of digital mammography One such efficiency found with digital mammography is the ability to perform remote reading. With telemammography, patients can be screened at one facility while the radiologist reads the image at a centralized site without transporting physical films. This means that one radiologist can service a variety of screening locations, thereby providing increased access without increasing the number of physicians. In addition, digital mammography offers a variety of other advantages. One is the immediacy of the process. The button is pushed and the image appears. The image can then be reviewed for quality control (QC) while the patient is still in the room, and the image can be retaken, if necessary. Another important advantage of digital mammography is its capacity to more clearly image dense tissue (Figure 1). The Digital Mammography Imaging Screening Trial from the American College of Radiology Imaging Network illustrated the value of this technology’s increased contrast resolution to better detect cancers in subgroups of women who predominantly have dense breasts.2 With digital mammography, the image acquisition is separate from display, leaving unlimited access to the original image. There also might be a reduction in the need for retakes because of incorrect technique with digital images. In the beginning, however, this may not be the case, as the technologist progresses through the learning curve on the digital system. Most importantly, however, digital data finally provides the potential to truly step into the next generation of breast imaging. Technologies that will provide a springboard for a digital platform are computer-aided diagnosis (CAD) and 3-dimensional imaging techniques, including tomographic imaging (such as tomosynthesis and computed tomography [CT]), subtraction, and dual-energy techniques. In theory, 1 digital system can replace 2 analog units, but in order to achieve this goal, workflow and connectivity must be streamlined. Seemingly simple tasks can take a long time to resolve, and, in our experience, our digital units have not been as reliable as our analog units. There are generally fewer steps for the technologists, less wait time for the patient, and faster throughput—as quick as 5 minutes of room time per study, without associated out-of-room tasks, which is the primary difference between analog and digital. Digital mammography is predicated on the fact that the technologist stays in, or at least very near to, the digital room. Efficiencies can best be realized with this approach. Digital mammography and the technologist Switching from analog to digital technology can seem like a daunting prospect for the technologist, but with a little patience, the technologist will become just as expert with digital mammography as with analog. It is not necessary to be computer literate to perform digital mammography; applications training will provide the necessary groundwork to use the equipment. New users should remain open-minded and spend time with the system. If your facility is switching to digital slowly, get in the digital room and use the system. The learning curve also includes the assimilation of proper terminology in order to converse in the language of digital mammography. For example, the terms “too light” and “too dark” are irrelevant terms, but “signal” and “noise” are most appropriate for the digital technology. The digital system In many ways, digital mammography is very similar to analog. In other ways it is quite different. Full-field digital mammography (FFDM) systems are composed of 3 main components: the acquisition stand or modality, the acquisition (or modality) workstation, and the diagnostic workstation (Figure 2). The modality or acquisition stand is very similar to that of an analog mammography system, with the defining difference being the detector. In a direct radiography (DR) FFDM system, a digital detector replaces the slotted bucky/cassette holder. The recently approved computed radiography (CR) mammography system employs digital cassettes similar to those used in analog systems. The modality workstation typically consists of a 1- or 2-megapixel (MP) monitor, a keyboard, and a computer. The technologist will perform acquisition-related tasks at the modality workstation, including setting technique, previewing images, and archiving or printing images. The diagnostic workstation, where the radiologist reads the image, consists of two 5-MP monitors for image review, a keypad to navigate the system, and, usually, a third nondiagnostic-grade monitor for workflow. When first approved, the FFDM systems were accompanied by a diagnostic workstation, but more recently, the U.S. Food and Drug Administration (FDA) has separated these components, leaving way for mammographic images to be read on a picture archiving and communication system (PACS) workstation. The PACS work-station, however, must be FDA-approved to display mammographic images. Using digital mammography systems The specifics of general digital technology are beyond the scope of this article; however, several very helpful articles on this topic have been published.3-5 In brief, with digital imaging, the X-ray beam or photon, now referred to as signal, is converted to an electronic or digital signal. As in general digital radiography, there are 2 types of digital mammography technology: direct radiography (DR) and CR (as of this writing, CR is nearing market approval for sale for mammography in the United States). With DR, there is no cassette. The breast support holds the digital receptor, and the image is captured directly on the unit and is ready for immediate display at the modality workstation. CR technology, on the other hand, employs cassettes that are similar to those used in analog units. Rather than taking film from the cassette to be developed, with CR the cassette is placed in a digital reader and the image is displayed at a workstation, which could also be located in the mammography room or may be shared between 2 X-ray rooms. Although digital image capture technology is significantly different from screen-film technology, the clinical technologist must still balance contrast and spatial resolution with dose and must apply appropriate techniques in order to obtain optimal image quality for interpretation. There are 2 aspects of mammography interpretation: detection and characterization. The radiologist first must be able to detect the lesion and then be able to characterize features in order to determine whether or not to perform a biopsy. For the technologist, it is important to learn how the digital system works and, when there is access to more than one mammography system, to know which system will provide the best image for the object being examined. Of course, technical application is greatly influenced by the interpreting radiologist, and the technologist should expect a learning curve for the radiologist as they and you learn to workup lesions differently. Processing The hallmark of digital imaging is that image acquisition, “processing,” and display are separated. When creating an image, the digital system first produces a raw data set, and then electronic processing is applied. The terminology for raw data is “for processing” and the processed image is known as “for presentation.” With early digital mammography systems, the acquisition stand provided the raw data and sent it to the diagnostic workstation to be processed. In the newer FFDM systems, the processing function has been moved to the acquisition stand. Processing algorithms, which are then applied to the “for processing” data, improve the presentation state of the image. While processing improves the display and makes the image more pleasing to the eye, processing cannot add information to the image. Currently, each FFDM vendor has its own proprietary processing formula, all of which are continuing to evolve. This will be the status quo for a few years, but in my opinion, we may be moving toward a system that employs processing boxes to which all raw data will be sent, regardless of origin, and all images will be processed in a consistent manner that will likely be chosen by the primary interpreting radiologist(s). Image acquisition and display Digital detector It is important to understand the origin of the digital image in order to properly apply technical factors. A digital acquisition platform is made up of small elements called pixels (or detector elements), which are arranged in a square or rectangular shape (Figure 3) referred to as a matrix. The space between the pixels is known as the pixel pitch. The size, pixel pitch, and arrangement of the pixels in the matrix provide for the spatial resolution of a system. Pixels also have bit depth (think of them as “wells” that fill up with X-ray photons). The bit depth of the pixel determines the number of shades of gray, which, in turn, provides the level of contrast resolution or—in digital speak—dynamic range. These factors are the essence of a digital receptor and, just like a screen-film combination, have characteristics that are unique to each detector. The FFDM systems in use today have detectors with pixel sizes ranging from 25 µm to 100 µm. As pixel size decreases, spatial resolution increases, but so do noise, radiation dose, and storage requirements. Looking at different matrices (Figure 4), one can see that, given the same bit depth, as the pixels get smaller and the number increases, the amount of information also increases. It is important to reiterate that as pixel size decreases, resolution will increase, but so will the noise and the dose factor. Manufacturers have carefully balanced pixel size and matrix in consideration of these factors. Technical applications Digital mammography has a new set of parameters for technical application. The technologist will no longer think in terms of mAs and kVp but in terms of signal and noise and the ratio between the two. Signal is the X-ray photons coming out of the tube. Noise is anything that interferes with the visibility of useful signal and includes quantum noise or mottle as well as electronic noise, a constant presence in digital detectors. Detective quantum efficiency (DQE), which is expressed as a percentage, is the ability of a system to detect and use exiting X-ray photons (the signal). In theory, the greater the DQE, the less signal (dose) is needed. The goal of technical application is to achieve adequate signal to fill the pixels and overcome existing electronic noise, without overfilling. While the beauty of a digital image is that display is separate from acquisition, this also adds difficulty in evaluating the quality of an image, as at first glance, a digital image will always look “good.” Under- and overexposure will not result in a “dark” or “light” image, but may result in a noisier image in the case of underexposure or poor contrast in the case of overexposure. Figure 5A shows an overexposed image, and Figure 5B shows the same image at the correct exposure; notice how flat in contrast the first overexposed image appears. This will not be able to be made better with windows and leveling. Note how more calcifications are apparent in the correctly exposed image. Note also the increase in noise in the “correctly” exposed image. The technologist will adjust mAs and kVp as a means to achieve adequate signal-to-noise ratio. Kilovoltage no longer has a great effect on image contrast but rather will be used to boost signal because image contrast is largely dependent on the dynamic range of the digital detector and the digital imaging chain. With digital technology, adequate exposure is measured with analog- to digital-units (ADUs) or exposure index (EI). Ranges for adequate exposure are provided by each manufacturer, and digital automatic exposure controls (AECs) allow for excellent exposure control. The technologist should understand that the entire detector is used for exposure control, in contrast to analog imaging, in which just a small, usually central phototiming detector was employed. Patient positioning The primary difference between DR and analog positioning is that the technologist has only one surface on which to position all breast and patient sizes. In addition, the digital detector is thicker, which means it is a little more difficult to accommodate a large abdomen and other more difficult body habitus. A positive aspect of the DR detectors is that the entire image receptor area acts as a digital AEC, which means that the breast does not have to be centered over a photocell. This allows a smaller breast to be positioned higher up on the image receptor for adequate positioning. Image display options Once the image is acquired and pro-cessed, it is typically displayed on monitors (however, in some situations, digital images may be printed to film for interpretation). Typically, the data is acquired at an acquisition matrix size of approximately 20,000,000 pixels in roughly a 10 × 12-inch matrix. The display (monitor) technology displays only 1 to 5 million pixels. What this means is that an entire image cannot be fully displayed on one monitor at full resolution. For this reason, softcopy workstations provide 3 ways for the images to be displayed. One option is called “fit to screen,” in which the information is downsized to fit to any window in which it is displayed. A second option is “true size,” which displays the image using the true size of the breast. Both options do not display the full data set of the digital image. A “pixel-to-pixel” or full-resolution display is the only display option that provides the entire data set. One issue that still needs to be addressed by FFDM manufacturers is the display monitor at the acquisition workstation. The acquisition workstation monitors are approximately 25 inches on the diagonal, with just 1 to 2 million pixels. This makes it more difficult for the technologist who must display the image in pixel-to-pixel mode and spend time panning the image to detect motion. Quality control Detecting motion At our facility, the technologists reported that it was difficult to detect motion on the 2-MP QC monitors in the imaging room. When we looked carefully at this issue, we found that the problem wasn’t in detecting motion, but more in confirming that there was no motion. On the 2-MP monitors, many images look somewhat fuzzy, although they appear smooth on the radiologists’ 5-MP monitor. One way to address this is to display the image using pixel-to-pixel resolution and check for gross motion. Artifacts As with analog imaging, artifacts can occur with digital mammography. Digital mammography artifacts are not yet widely understood and can be a time-consuming problem to resolve. Some artifacts may be very subtle, but they can have a dramatic effect. Artifacts arise from the detector, processing, and the monitor as well as the X-ray tube, filter, and grid. The radiologist has a higher-resolution monitor and may see artifacts more readily than the technologist. Monitor artifacts can be particularly difficult because the technologists and the radiologists are viewing the image on different monitors. Therefore, an artifact on the radiologists’ 5-MP monitor will not appear on the 2-MP monitor that the technologist used to perform the QC. DR systems directly convert X-ray to digital signal. Such units may exhibit trouble in completely clearing the imaging detector of the previous image or images, which can result in “ghosting” artifacts (Figure 6). It is not clear at what level ghosting noise interferes with the diagnostic quality of an image, if at all. The image displayed in Figure 6 was acquired using the QC flat-field phantom. Outlines of previous mammogram images are visible as a ghosting artifact. The ghosting may get worse with detector age, as was the case with this early version of a DR detector, which had to be replaced. Quality assurance Quality assurance measures for digital mammography are still evolving. Currently, system testing is conducted using manufacturing paradigms that are approved by the Mammography Quality Standards Act (MQSA) for each individual vendor’s system, but the American College of Radiology (ACR) is working on developing a uniform set of standard tests for all systems. All parts of the digital unit must undergo quality assurance, including the diagnostic workstation, the acquisition workstation, the acquisition stand, and the printer. System servicing Servicing of FFDM systems is vastly different from that of analog systems, and this has both advantages and disadvantages. Some servicing issues can be worked out over the phone, thereby eliminating downtime while awaiting the arrival of a service technician. The downside to this is that the technologist, in effect, becomes the service person, often working in tandem with a technician on the phone. Therefore, it is very important to have a phone, preferably a cordless one or one with a very long cord, in the digital room. Another advantage to servicing an FFDM system is the ability to conduct remote dial-in servicing. This means that the servicing company can sometimes diagnose and rectify a problem remotely through a modem or virtual private network. Field service engineers and technologists are on a learning curve for troubleshooting, and the technologist and service person may need to work together to determine the cause of a problem. Historically, service engineers have been mechanically oriented. Now, with the new technology, they must be more computer savvy. If you have the first digital system in your area, it’s likely that you’re going to be on the learning curve for that service engineer. Servicing and performing quality assurance testing of FFDM systems may also be a matter of geography. The diagnostic workstation will be remote to the acquisition stand, and a printer may be in a third location, yet all of these components must be tested, maintained, and documented according to MQSA standards. Conclusion The promise of digital mammography lies not in the use of the abundant data set to produce a 2-dimensional image, but in the ability of FFDM to take us to the next generation of breast imaging, including tomosynthesis, CT mammography, contrast-enhanced subtraction mammography, CAD, and, eventually, less invasive methods of biopsy and treatment of breast cancer, blurring the lines between diagnosis and treatment. For the technologist, however, although the tools may change, their primary job is that of listening to and forming a relationship with the patient, in order to provide the best images and pertinent information for interpretation. That does not, and will not, change with digital mammography or any other new technology. REFERENCES |