NIST GCR 03-844
Low-Cost Manufacturing Process Technology for Amorphous Silicon Detector Panels: Applications in Digital Mammography and Radiography

3. Digital Mammography

Cancer is a group of diseases characterized by uncontrolled growth and spread of abnormal cells. It is caused by genetic and environmental factors. Ten or more years may pass between the occurrence of mutations or exposure to environmental factors and detectable cancers.

Breast cancer is a devastating disease and is the second most common cause of cancer-related death in the United States. In 2002, approximately 205,000 new cases of invasive breast cancer will be diagnosed, and about 40,000 women are expected to die from the disease (American Cancer Society, 2002a).

Conventional film-based mammography is a screening technology for the early detection of breast cancers. Early detection is vital to facilitating more effective treatment. The new technology of digital mammography was recently developed to go beyond the clinical capabilities of conventional film mammography and to address its productivity limitations.

BREAST CANCER AND EARLY DETECTION

Breast cancer is a complex of more than twenty distinct types of breast disease (Thompson, 2002). It progresses in stages. The first, in situ stage is limited to small areas and confined to the cells in which the cancer began, without invasion of surrounding tissues. The second, localized stage is invasive cancer into surrounding mammary tissue, but not invasive outside the breast tissue. In the third, regional stage, cancer spreads to the chest wall, muscles, and immediately upstream lymph nodes. In the fourth or distant stage, cancerous tumor cells have metastasized or spread to other parts of the body.

Early detection, prior to cancer spreading from initial cells to adjacent breast tissue and other regions of the body, is widely believed to save lives by facilitating early intervention, when treatment is most likely to be effective. If detected “when small and local, treatment options may be less dangerous, intrusive, and costly and morelikely to lead to a cure” (Institute of Medicine, 2001). At the same time, in clinical practice there is an exception to every rule and early detection of some slow-growing cancers could lead to overly aggressive treatment causing unnecessary medical risks and treatment expenses. But because medicine is an applied science of probabilities rather than certainties, for the majority of cases early detection is widely considered to be a desirable outcome.

BREAST CANCER TRENDS

Breast cancer primarily affects female populations and is relatively uncommon in men. The National Cancer Institute estimates that about one in eight, or 13 percent, of women in the United States will develop breast cancer during their lifetime.

The risk of developing breast cancer increases with age (Table 2). More than 80 percent of breast cancers occur in women aged fifty years or older. It is uncommon under the age of forty.

Age-related incidence rates do not properly characterize high-risk populations. Five to seven percent of breast cancer is hereditary, linked to BRCA1, BRCA2, and other genes. These mutations carry lifetime risks on the order of 56 to 85 percent, compared to 13 percent average lifetime risk. Additional groups of high-risk populations include women with a family history of breast cancer, personal cancer history, and possibly higher breast density (Jardines, 2001). Obesity and urban residence also correlate with higher risk (Miller, 1996).

Incidence patterns vary by race and ethnicity, “revealing that white, Hawaiian, and black women have the highest age adjusted rates. Lowest U.S. rates occur among Korean, American Indian, and Vietnamese women” (Miller, 1996). Mortality patterns by race and ethnicity differ from incidence patterns. “The highest age adjusted mortality occurs among black women, followed by white and Hawaiian women. Higher mortality among black women is thought to be related to a larger percentage of breast cancers being diagnosed at a later, less treatable stage” (Miller, 1996).

During the 1987–1999 period there was an increase in the incidence of breast cancer. This trend is expected to continue as “the U.S. population grows older and cancer is largely a disease of older people.” At the same time, “breast cancer death rates are down, due to screening mammography, which catches cancers earlier when they are more treatable, and to improved treatment modalities” (American Cancer Society, 2002b).

Table 2. Incidence of Breast Cancer for American

CONVENTIONAL MAMMOGRAPHY

Mammography is a medical imaging procedure for breast cancer screening and diagnosis. Screening mammography is the X-ray imaging of the breasts of women with no complaints or symptoms of breast cancer. The goal is to detect asymptomatic cancer when it is still too small to be felt by a woman or a physician. Diagnostic mammography is an X-ray examination of the breasts of women who either have a breast complaint (for example, a breast lump or nipple discharge) or when an abnormality has been found during previous screening mammography. Eighty percent of U.S. mammography procedures are used for cancer screening.

Conventional screen film mammography (SFM) uses a low-dose X-ray system andhigh-contrast, high-resolution film to create detailed images of the breasts. To take a mammogram, the X-ray source is turned on and low dose X-rays are radiated throuh the compressed breast and onto a film cassette positioned under the breast. SFM procedures consist of X-ray production, differential X-ray absorption by breast tissues, recording of transmitted X-rays on photographic film, film development, and image display for reading or interpretation by a specialized radiologist (mammographer).

Breast abnormalities or lesions that may be detected through mammography include calcification and lumpy masses. The goal of the mammographer is to detect those abnormalities that can be suggestive of malignancy.

• Calcifications are tiny mineral deposits within the breast tissue that appear as small white spots on films. Macrocalcifications are coarse (larger) mineral deposits that most likely represent degenerative changes in the breast, such as aging of arteries, injuries, or inflammations. These deposits are usually associated with benign (noncancerous) conditions and do not require a subsequent biopsy. Macrocalcifications are found in about 50 percent of women over the age of fifty. Microcalcifications are tiny specs of calcium in the breasts, may appear singly or in clusters, and may be indicative of cancer. The shape and arrangement of microcalcifications help the radiologist judge the likelihood of cancer being present and the need to prescribe a biopsy (American Cancer Society, 2002c).
• Masses are abnormal breast tissue that may be benign or cancerous. Some masses are fluid-filled spaces called cysts. To evaluate cysts, the radiologist may prescribe a breast ultrasound procedure. Sometimes radiologists will decide to aspirate the cyst, that is, remove a fluid sample from the cyst using a needle. The fluid is then examined by a pathologist to determine if the cells are cancerous. For masses other than cysts, the size, shape, and edges of the mass may be indicative of cancer and the radiologist could prescribe a biopsy.

For the most effective interpretation of mammograms and the detection of calcifications and lumpy masses, it is important to have prior mammograms available for comparison. If prior mammograms indicate that calcifications and mass patterns have not changed, benign condition may be indicated and unnecessary diagnostic procedures avoided.

COMPLEMENTARY BREAST SCREENING MODALITIES

Other breast screening modalities (technologies) that complement mammography include:

• Ultrasound. High-frequency sound waves are bounced off breast tissues. The echoes produce a picture called a sonogram. Ultrasound imaging can be used to distinguish between solid tumors and fluid-filled cysts and evaluate lumps that are hard to see on mammograms. Ultrasound is not used for routine breast cancer screening as it does not consistently detect early signs of cancer such as microcalcifications.
• Magnetic Resonance Imaging. A magnet linked to a computer creates a series of detailed cross-sectional pictures of the breast without the use of radiation. Like ultrasound, MRI cannot detect microcalcifications and is generally not used for cancer screening except in special cases, such as screening high-risk women with a strong family history of breast cancer, where “MRI may be a superior screening modality” (Stoutjesdijk, 2001).
• Computed Tomography. A pencil-thin beam of high-energy radiation is used to create a series of breast images, taken from different angles. The images are fed into a computer and combined into a single image of the breast.

In general, the above technologies are used to complement and not displace mammography as the primary modality for breast cancer screening. “Though imperfect, mammography remains the best screening tool as well as the gold standard for breast cancer diagnosis” (Institute of Medicine, 2001).

SCREENING MAMMOGRAPHY TRENDS

The use of screening mammography for detecting breast cancer has increased dramatically. In 1987, 29 percent of women over forty years of age had a mammogram in the past two years. In 1998, more than 60 percent had a mammogram. As Figure 2 indicates, expanded mammography use has reached most racial and ethnic groups, with only Hispanic women lagging average use rates by 6 percent. In 2000, more than 32.5 million mammograms were performed in the United States. Further increases in screening mammography use are expected to encounter substantial barriers, including:

• High volume workloads coupled with a growing shortage of radiologists.
• High mammography costs coupled with lack of medical insurance.
• Poor accessibility in some isolated, rural, or inner city neighborhoods, and for the growing population of older women (Lichtman, 1996).

LIMITATIONS OF SCREENING MAMMOGRAPHY

Since the 1960s, eight clinical trials of screening mammography have been conducted in the United States, Sweden, the United Kingdom, and Canada. The results of these clinical trials indicated substantial reductions (more than 30 percent) in breast cancer mortality and were extensively used to justify national screening mammography programs.

In 2001, the Nordic Cochrane Centre commissioned a meta-analysis of the results of these clinical trials. The authors of the meta-analysis argued that the clinical trials were not properly randomized and concluded “that there is no reliable evidence that screening mammography reduces breast cancer mortality and breast cancer screening may not be worth the physical and financial toll it exerts.” Radiation exposure was also a concern (Gotzsche and Olsen, 2001).

A subsequent, February 2002, “fresh look” at the original data of the Swedish clinical trials concluded that mammography does save lives, but at a somewhat lesser rate than claimed in the original clinical studies (Nystrom et al., 2002). The predominant U.S. and international medical opinion largely accepts the Nystrom rebuttal of the Cochrane meta-analysis concerning the value of screening mammography.

• According to the American Cancer Society, “the overwhelming weight of evidence shows that mammography saves lives” (Norton, 2002).
• According to the International Agency for Research on Cancer, “a woman who is screened regularly can expect about 35 percent reduction in her risk of death from breast cancer” (Reaney, 2002).
• According to the American College of Radiology, “there is sufficient data that clearly demonstrates that screening mammograms are saving lives” (AuntMinnie, 2002).

While medical opinion is generally supportive of screening mammography, there is a parallel tendency to recognize its limitations. These limitations revolve around the subjectivity of image interpretation.

Figure 2. Percentage of Women (Aged 40+) Who Had Mammograms Within Past 2 Years, by Race and Ethnicity

In particular, screening mammography results are currently characterized by:

• Twenty-percent rate of false negatives where mammography fails to detect cancerous conditions (National Cancer Institute, 2002).
• Inadequate screening of high-risk women, necessitating more frequent examinations (beyond annual) and other screening modalities such as MRIs (Stoutjesdijk et al., 2001).
• Missing rapidly proliferating, high-grade tumors or detecting these too late (Porter, 1999).
• Tendency to miss cancers in women with dense breast tissue. The breasts of younger women contain many glands and ligaments that appear dense on a mammogram making it difficult to spot tumors. High-quality mammograms detect approximately 90 percent of cancers in women over fifty but only 60 percent of breast cancers in women under fifty (National Cancer Institute, 2002).
• Up to 12 percent false-positive rate where incorrect interpretation identifies cancer where none is present. False-positive readings result in unnecessary recalls for diagnostic mammography and biopsies, unnecessary medical and transportation expenses, and unnecessary anxiety and physical discomfort (Alexander, 1999).
• Concern about radiation exposure.

Longitudinal comparison of mammograms can alleviate some of the above limitations and is frequently used by radiologists to distinguish microcalcifications and breast lumps that are likely to be malignant (Medscape, 2001). However, at typical U.S. imaging centers, 20 percent of archived film is lost or otherwise unavailable for clinical comparison, impeding the analytical continuity of mammogram interpretation.

DIGITAL MAMMOGRAPHY

In digital mammography, a digital detector is added to replace the film cassette of the conventional system (Figure 3). Due to the detector’s high detective efficiency, or detective quantum efficiency (DQE), it has the potential to capture up to 82 percent of the original signal or breast image information. Digital mammography also operates at significantly higher speeds, facilitates the independent interpretation of mammograms by two radiologists (double reading), and supports the development of regional telemammography networks, designed to reach isolated and underserved populations.

Digital mammography uses electronic detector panels for capturing the X-ray images of the breast as a collection of discrete electrical charges. After the X-ray passes through the tissue undergoing imaging, it encounters the scintillator layer of the detector that converts X-rays to light. Next, light signals encounter the detector’s photodiode layer where signals are converted into electrical charges. Electrical charges are converted to voltage signals and transferred to a monitor for image display and to a computer for image analysis and storage.

GE uses a-Si for the fabrication of the detector photodiode layer. A-Si makes it possible to fabricate large panels of one continuous piece for capturing a full breast in one high-resolution image. This capability is referred to as full-field digital mammography (FFDM) and is deemed to be desirable for optimal mammogram reading and interpretation. A-Si appears to be the most promising design for doseefficient FFDM, as well as for other high performance medical applications (Granfors and Aufrichtig, 2000). Alternative design approaches may be appropriate for less demanding imaging tasks.

Figure 3. Direct Digital Mammography

“Though imperfect, screen filmmammography remains the best screening tool for breast cancer detection” (Institute of Medicine, 2001). The new technology of digital mammography was recently developed to go beyond the clinical capabilities of conventional film mammography and to effectively address its limitations.

CLINICAL TRIALS: DIGITAL MAMMOGRAPHY

A recently completed clinical trial funded by the U.S. Army Medical Research and Material Command and conducted through the University of Colorado and University of Massachusetts Medical Centers compared the screening accuracy of FFDM and conventional film mammography on 5,000 subjects. The study identified no statistically significant difference in cancer detection. However, digital mammography had a 20 percent lower recall rate from false positive readings (Lewin et al., 2000).

The Army clinical trial used a relatively small sample size, and “used prototype digital systems with lower resolution, lower luminance monitors, and inferior image processing capabilities” compared to the Senographe 2000D, the current FDAapproved GE FFDM (Schubert, 2002).

Given the limitations of the U.S. Army study as well as technical advances since the prototype stage, the radiological community generally anticipates improved results from the new FDA-approved fully commercial FFDM system. Per the National Cancer Institute, “digital mammography has the potential to provide better detection of early breast cancer, but a large study is needed to really determine whether it is better than conventional mammography and how large the difference” (RSNA, 2002). To identify whether the anticipated performance improvements are being realized and to generate results with higher statistical validity, the National Cancer Institute and the American College of Radiology Imaging Network (ACRIN) have initiated a large, multi-center clinical trial using various digital mammography devices made by four different manufacturers (GE, Fuji Medical Systems, Fischer Imaging, and LORAD). The $26.3 million clinical trial with a screening population of 49,500 women is currently underway in the United States and Canada (Pisano, 2001).

COMPUTER-AIDED DETECTION

FFDM facilitates the use of computer-aided detection (CAD), a process in which radiologists use computers and specialized software to read mammograms and identify breast abnormalities. “Currently, mammograms are visually examined in search of subtle and complicated indicators of breast cancer. This can be a difficult, tedious, and time-consuming task as only one in a thousand mammograms may show an abnormality of concern. Computers can assist mammographers by consistently scanning every part of every mammogram and reporting suspicious areas. This allows the human expert to make the most efficient use of his/her time and focus on those cases generating the greatest concern” (Lawrence Livermore, 1995).

A 12-month study of 12,900 women, screened with mammography plus CAD, identified a 19.5 percent increase in the number of cancers detected relative to mammography without CAD (Freer, 2001). Another smaller 2002 study identified a more than 50 percent increase (Jong et al., 2002). Other studies pointed to the benefits of CAD in reducing variability in radiologist interpretations (Yulei et al., 2001) and the greater sensitivity of CAD for breast masses as opposed to calcifications (Markey et al., 2002).

POTENTIAL BENEFITS FROM DIGITAL MAMMOGRAPHY

The ongoing ACRIN clinical trials may find superior cancer detection performance over conventional film mammography. However, even in the absence of findings indicating improved cancer detection, digital mammography can be associated with the following medical, economic, and social benefits that make it an attractive screening tool for diagnostic imaging centers.

Medical Benefits

• Lower false positive rates, and therefore fewer unnecessary biopsies.
• Lower call-back rates for mammogram over- and under-exposure, and therefore avoidance of unnecessary procedures.
• Reduced radiation exposure for women with dense breast tissue.
• Simplified retrieval, and elimination of loss, of prior mammograms, facilitating analytical continuity and improved cancer detection.
• Facilitation of use of CAD for improved cancer detection.
• Enhanced real-time sharing of mammograms in clinical settings for double reading, associated with improved cancer detection.

Productivity and Economic Benefits

Increased throughput (counteracting the growing shortage of radiologists), reduced patient examination time, and reduced waiting time.

• Reduced lost wages and travel time for patients by avoiding unnecessary recalls.
• Reduced medical diagnostic costs from avoided recalls and biopsies.
• Reduced operating costs by eliminating film and film development.
• Reduced record-keeping costs by eliminating film archiving.

Social Benefits

Facilitation of regional telemammography networks, thus expanding access to quality mammography services at underserved rural locations (for example, at military bases and field hospitals) and by underserved minority populations.

Return to Table of Contents or go to Section 4. Digital Chest Radiography and Other Applications.

Date created: April 25, 2003
Last updated: August 2, 2005

ATP website comments: webmaster-atp@nist.gov  / Technical ATP inquiries: InfoCoord.ATP@nist.gov.

NIST is an agency of the U.S. Commerce Department
Privacy policy / Security Notice / Accessibility Statement / Disclaimer / Freedom of Information Act (FOIA) /
No Fear Act Policy / NIST Information Quallity Standards / ExpectMore.gov (performance of federal programs)