**Digital Mammography**

Cherie M. Kuzmiak *University of North Carolina USA* 

#### **1. Introduction**

80 Imaging of the Breast – Technical Aspects and Clinical Implication

Van Eijck, CH; Krenning, EP; Bootsma, A; Oei, HY; Van Pel, R; Lindemans, J; Jeekel, J;

Waxman, AD; Ramanna, L; Brachman, MB. (1989). Thallium scintigraphy in primary

breast cancer. *Lancet,* Vol.343, (1994), pp.640–643, ISSN 0140-6736

Vol.30, (1989), pp.844–848, ISSN 0161-5505

Reubi, JC; Lamberts, SWJ. (1994). Somatostatin-receptor scintigraphy in primary

carcinoma of the breast: evaluation of primary and axillary metastasis. *J Nucl Med,* 

Full-field digital mammography has transformed mammography over the past decade. The technology has reached a level of maturity that has caused an increase in its utilization in hospitals and clinics world-wide. This chapter will discuss the advantages and disadvantages of the technology as compared to screen-film mammography, a discussion on the basic physics of digital mammography and the currently available detector technologies. In addition to the technical aspects, this chapter will explore the clinical trials published to date regarding the technology performed compared to screen-film mammography in both screening and diagnosis, and evaluate the various imaging process algorithms that have been applied to digital mammography. Finally, digital mammography's impact on daily clinical workflow will be discussed along with future directions for this technology.

Digital mammography has become part of everyday clinical practice across much of the developed world. However, I find it interesting that most of the radiologists in training (residents and fellows) have no concept of what "film" is and was to our practice of mammography. They are used to the digital age of computers, personal electronic devices and electronic social networking through the internet. To the present generation, softcopy display in some form is a way of daily life. To others who have devoted their medical career in screening and in the diagnostic evaluation of women for breast cancer, it has and continues to be a learning and transitional process. In order to understand and fully appreciate the advances in mammography technology, it is important to understand the natural history of breast cancer and the challenges and changes that our specialty has undergone and how it continues to evolve.

Cancer of the breast is not a new disease. It has been present since ancient times and was documented by the early Egyptians, Greeks, Babylonians and Chinese (Bland, 1998). If a woman presented to her local "healer/physician" with a lump in her breast, treatment may have stemmed from charms and chants to applied ointments or possibly intervention with a knife and hot irons for cauterization. For women, treatment for breast cancer was similar for centuries and prognosis was generally poor.

It was not until 1913, when Albert Salomon, a German surgeon, evaluated 3,000 mastectomy specimens in a radiology-histological study on comparing the x-ray findings with microscopic pathology that it became evident to evaluate radiography technology for breast cancer detection. In the 1920's and 30's, several attempts to implement radiography for the diagnosis of breast abnormalities were done by O. Kleinschmit, W. Vogel, J. Goyanes, and

Digital Mammography 83

Fig. 1. Hurter and Driffield (H&D) curve for a SFM system where Log E represents the

With the limited range of soft-tissue densities in the breast, mammography requires high contrast. The fixed characteristics of an H&D curve mean that if high contrast is to be obtained in intermediate-density tissue, there must be lower contrast within the thicker, denser fibroglandular tissues represented at the toe of the curve and the fatty tissue represented at the shoulder of the curve (Feig & Yaffe, 1998). Consequently, mammographic

**Log E**

Digital mammography offers several advantages over SFM. The digital system separates the process of x-ray detection from image display and storage. Since image acquisition and display are separated, each can be optimized. Digital detectors have a wider dynamic range (linear response) compared to film as seen in Figure 2. Digital detectors have increased efficiency at a lower radiation dose in the detection and depiction of the x-ray photons compared to film (Pisano, 1998; Feig, 1996). In addition, digital detectors (even with a lower spatial resolution than film) also appear to improve lesion conspicuity through their improved efficiency of absorption of x-ray photons, a linear response over a wide range of radiation intensities and low system noise (Feig & Yaffe, 1998). Plus, post-processing software can be utilized to assist the radiologist in evaluating the images for suspicious findings by altering contrast and brightness automatically or manually. With digital mammography, computer aided detection software can be utilized at a push of a button instead of waiting for someone to digitize the film images for each case. With digital

mammography, the images can be displayed with hard and softcopy formats.

relative exposure and OD is the optical density.

film has a limited dynamic range.

**Response - OD**

Gershon-Cohen; however, it was not until four decades later that the use of x-rays for the diagnosis of breast became more established (Picard, 1998).

The modern era of mammography began in the late 1960's as the technique was refined with dedicated equipment, such as that developed by the physicist C. Gros (Van Steen & Van Tiggelen, 2007). During this time, the film industry began to develop dedicated mammography film with high-quality images, reliable capture parameters and reduced radiation dose to the patient. Previously, a general x-ray tube with industrial film (low sensitivity) with high radiation exposure to the patient was used to image the breast. By the late 1970's and 80's, dedicated mammography was established, and mammography was identified as the most reproducible and cost effective modality to screen the general population. In clinical studies with follow-up of patients, it was shown that early detection of breast cancer has led to a reduction of the mortality rate (range 18-30%) (Elmore, 2005; Hendrick, 1997; Nystrom, 1993; Strax, 1973; Tabar, 2011).

It is estimated that 1.38 million women were diagnosed with breast cancer worldwide in 2008. This accounted for approximately a tenth (10.9%) of all the new cancers and 23% of all female cancers (Ferlay, 2010). Female breast cancer rates vary. The highest rates are in Europe and the United States and the lowest are currently in Africa and Asia. Currently, breast cancer is the second leading cause of death in the women of the United States (Center of Disease Control, 2010) and the United Kingdom (Ferlay, 2010). Because of these cancerrelated deaths and the continued incidence of the disease worldwide, further emphasis has been placed on using mammography as screening tool for early detection.

#### **2. Physics of digital mammography**

#### **2.1 Comparison of screen film mammography**

Screen-film mammography (SFM) has been (and continues to be in some countries) the standard imaging modality for detecting suspicious lesions at an early stage in the breasts of asymptomatic women. Film is a very useful medium that has been optimized over the past 50 years. SFM has a high sensitivity (100%) in detecting suspicious lesions in breasts composed primarily of fatty tissue (Dujm, 1997; Saarenmaa, 2000). However, that value is significantly decreased in breasts composed of dense glandular tissue because breast cancers are frequently similar in radiographic density to the fibroglandular tissue. Consequently, 10- 20% of breast cancers are not visualized (Burrell et al., 1996). Also, part of this decrease in lesion conspicuity may be due to the film itself since it serves as the medium of image acquisition, display and storage. After the film is exposed and processed, the image cannot be significantly altered and portions of the mammogram may be displayed with suboptimal contrast. Only slight improvements can be made with a "hot light" or magnifying glass. If improvements cannot be made, the patient may need to undergo another mammographic image and consequently be exposed to more radiation dose.

Another limitation of film is that different regions of the breast image are represented according to the characteristic response of the mammographic film. There is a trade-off between the dynamic range (latitude) and contrast resolution (gradient). This is illustrated in Figure 1 by the sigmoid Hurter and Driffield (H&D) curve that is characteristic for a given type of SFM system under specific conditions. The H&D curve demonstrates the relationship between x-ray exposure, image density and contrast (Feig & Yaffe, 1998).

Gershon-Cohen; however, it was not until four decades later that the use of x-rays for the

The modern era of mammography began in the late 1960's as the technique was refined with dedicated equipment, such as that developed by the physicist C. Gros (Van Steen & Van Tiggelen, 2007). During this time, the film industry began to develop dedicated mammography film with high-quality images, reliable capture parameters and reduced radiation dose to the patient. Previously, a general x-ray tube with industrial film (low sensitivity) with high radiation exposure to the patient was used to image the breast. By the late 1970's and 80's, dedicated mammography was established, and mammography was identified as the most reproducible and cost effective modality to screen the general population. In clinical studies with follow-up of patients, it was shown that early detection of breast cancer has led to a reduction of the mortality rate (range 18-30%) (Elmore, 2005;

It is estimated that 1.38 million women were diagnosed with breast cancer worldwide in 2008. This accounted for approximately a tenth (10.9%) of all the new cancers and 23% of all female cancers (Ferlay, 2010). Female breast cancer rates vary. The highest rates are in Europe and the United States and the lowest are currently in Africa and Asia. Currently, breast cancer is the second leading cause of death in the women of the United States (Center of Disease Control, 2010) and the United Kingdom (Ferlay, 2010). Because of these cancerrelated deaths and the continued incidence of the disease worldwide, further emphasis has

Screen-film mammography (SFM) has been (and continues to be in some countries) the standard imaging modality for detecting suspicious lesions at an early stage in the breasts of asymptomatic women. Film is a very useful medium that has been optimized over the past 50 years. SFM has a high sensitivity (100%) in detecting suspicious lesions in breasts composed primarily of fatty tissue (Dujm, 1997; Saarenmaa, 2000). However, that value is significantly decreased in breasts composed of dense glandular tissue because breast cancers are frequently similar in radiographic density to the fibroglandular tissue. Consequently, 10- 20% of breast cancers are not visualized (Burrell et al., 1996). Also, part of this decrease in lesion conspicuity may be due to the film itself since it serves as the medium of image acquisition, display and storage. After the film is exposed and processed, the image cannot be significantly altered and portions of the mammogram may be displayed with suboptimal contrast. Only slight improvements can be made with a "hot light" or magnifying glass. If improvements cannot be made, the patient may need to undergo another mammographic

Another limitation of film is that different regions of the breast image are represented according to the characteristic response of the mammographic film. There is a trade-off between the dynamic range (latitude) and contrast resolution (gradient). This is illustrated in Figure 1 by the sigmoid Hurter and Driffield (H&D) curve that is characteristic for a given type of SFM system under specific conditions. The H&D curve demonstrates the relationship between x-ray exposure, image density and contrast (Feig & Yaffe, 1998).

been placed on using mammography as screening tool for early detection.

diagnosis of breast became more established (Picard, 1998).

Hendrick, 1997; Nystrom, 1993; Strax, 1973; Tabar, 2011).

**2. Physics of digital mammography** 

**2.1 Comparison of screen film mammography** 

image and consequently be exposed to more radiation dose.

Fig. 1. Hurter and Driffield (H&D) curve for a SFM system where Log E represents the relative exposure and OD is the optical density.

With the limited range of soft-tissue densities in the breast, mammography requires high contrast. The fixed characteristics of an H&D curve mean that if high contrast is to be obtained in intermediate-density tissue, there must be lower contrast within the thicker, denser fibroglandular tissues represented at the toe of the curve and the fatty tissue represented at the shoulder of the curve (Feig & Yaffe, 1998). Consequently, mammographic film has a limited dynamic range.

Digital mammography offers several advantages over SFM. The digital system separates the process of x-ray detection from image display and storage. Since image acquisition and display are separated, each can be optimized. Digital detectors have a wider dynamic range (linear response) compared to film as seen in Figure 2. Digital detectors have increased efficiency at a lower radiation dose in the detection and depiction of the x-ray photons compared to film (Pisano, 1998; Feig, 1996). In addition, digital detectors (even with a lower spatial resolution than film) also appear to improve lesion conspicuity through their improved efficiency of absorption of x-ray photons, a linear response over a wide range of radiation intensities and low system noise (Feig & Yaffe, 1998). Plus, post-processing software can be utilized to assist the radiologist in evaluating the images for suspicious findings by altering contrast and brightness automatically or manually. With digital mammography, computer aided detection software can be utilized at a push of a button instead of waiting for someone to digitize the film images for each case. With digital mammography, the images can be displayed with hard and softcopy formats.

Digital Mammography 85

digitized and stored. With digital mammography, wet chemical processing is eliminated and the detector's only role is image acquisition. Another added benefit of a digital detector

Spatial resolution in SFM is commonly based on the limiting resolution in terms of linepairs/mm from a bar pattern, Figure 3. This test can be very subjective. Therefore, in order to evaluate spatial resolution more quantitatively, it can be evaluated with the modulation transfer function (MTF). The MTF describes how well the entire imaging system or one of its components is performing in the form of a sinusoidal shape (Bunch, 1987; Pisano, 2004). The MTF describes how well each spatial frequency is transferred through a system. The MTF of a system is the product of the MTFs of the components of each system. As seen in Figure 4, at low spatial frequencies the MTF value is at or near the value of 1.0 and the MTF value decreases with increasing spatial frequency. The MTF of SFM extends beyond 20 cycles/mm and it is predominately the result of the screen since film has a very high MTF (Pisano et al., 2004). In a digital system, the MTF will be based on the focal spot, patient motion, lateral

spread of signal (light or electronic charges) in the detector, and spatial sampling.

Fig. 3. Bar pattern of line-pairs/mm for determining spatial resolution for SFM.

Fig. 4. Modulation Transfer Function (MTF) of a screen film system.

is the elimination of film granularity that adds noise to a system.

**2.3 Properties of digital images** 

Fig. 2. Digital detectors have a linear response and wide dynamic range compared to SFM. The digital response is seen as the diagonal line. Log E represents the relative exposure and OD is the optical density.

#### **2.2 Image acquisition**

To obtain a mammographic image, x-rays must be generated from a target. A metal filter in the system will remove the majority of non-desirable energies of the beam before it enters the patient. In a SFM system, the automatic exposure control (AEC) will end the film exposure when the tissue above the AEC has transmitted a suitable number of x-rays to expose the film where its gradient (slope of the H&D curve) will be near or at its maximum value and there will be acceptable image brightness (Yaffe, in Bick & Diekmann, 2010). However, other areas of the breast may be suboptimally exposed - dense areas underexposed. In the SFM system, the intensifying screen produces light that is proportional to the amount of energy deposited by the x-rays. The now exposed film will be chemically processed to produce the permanent mammographic image of the different optical densities. The mammographic film serves the three roles of image acquisition, display and storage.

In digital mammography, a detector replaces the screen-film system. The detector will still be exposed to x-rays just as in a SFM system. The detector produces a signal that is linearly proportional to the intensity of the photons transmitted by the breast; therefore, it is possible to produce a better representation of the x-ray transmission of all parts of the breast (Feig & Yaffe, 1998). For digital mammography, AEC still plays a role. Unlike SFM where it was important in determining image contrast, in digital mammography the AEC aids in obtaining a predetermined signal-to-noise ratio and a reasonable radiation dose to the breast. After being exposed, the digital detector produces an electronic signal that is

Fig. 2. Digital detectors have a linear response and wide dynamic range compared to SFM. The digital response is seen as the diagonal line. Log E represents the relative exposure and

**Log E**

To obtain a mammographic image, x-rays must be generated from a target. A metal filter in the system will remove the majority of non-desirable energies of the beam before it enters the patient. In a SFM system, the automatic exposure control (AEC) will end the film exposure when the tissue above the AEC has transmitted a suitable number of x-rays to expose the film where its gradient (slope of the H&D curve) will be near or at its maximum value and there will be acceptable image brightness (Yaffe, in Bick & Diekmann, 2010). However, other areas of the breast may be suboptimally exposed - dense areas underexposed. In the SFM system, the intensifying screen produces light that is proportional to the amount of energy deposited by the x-rays. The now exposed film will be chemically processed to produce the permanent mammographic image of the different optical densities. The mammographic film serves the three roles of image acquisition,

In digital mammography, a detector replaces the screen-film system. The detector will still be exposed to x-rays just as in a SFM system. The detector produces a signal that is linearly proportional to the intensity of the photons transmitted by the breast; therefore, it is possible to produce a better representation of the x-ray transmission of all parts of the breast (Feig & Yaffe, 1998). For digital mammography, AEC still plays a role. Unlike SFM where it was important in determining image contrast, in digital mammography the AEC aids in obtaining a predetermined signal-to-noise ratio and a reasonable radiation dose to the breast. After being exposed, the digital detector produces an electronic signal that is

OD is the optical density.

**Response - OD**

**2.2 Image acquisition** 

display and storage.

digitized and stored. With digital mammography, wet chemical processing is eliminated and the detector's only role is image acquisition. Another added benefit of a digital detector is the elimination of film granularity that adds noise to a system.

#### **2.3 Properties of digital images**

Spatial resolution in SFM is commonly based on the limiting resolution in terms of linepairs/mm from a bar pattern, Figure 3. This test can be very subjective. Therefore, in order to evaluate spatial resolution more quantitatively, it can be evaluated with the modulation transfer function (MTF). The MTF describes how well the entire imaging system or one of its components is performing in the form of a sinusoidal shape (Bunch, 1987; Pisano, 2004). The MTF describes how well each spatial frequency is transferred through a system. The MTF of a system is the product of the MTFs of the components of each system. As seen in Figure 4, at low spatial frequencies the MTF value is at or near the value of 1.0 and the MTF value decreases with increasing spatial frequency. The MTF of SFM extends beyond 20 cycles/mm and it is predominately the result of the screen since film has a very high MTF (Pisano et al., 2004). In a digital system, the MTF will be based on the focal spot, patient motion, lateral spread of signal (light or electronic charges) in the detector, and spatial sampling.

Fig. 3. Bar pattern of line-pairs/mm for determining spatial resolution for SFM.

Fig. 4. Modulation Transfer Function (MTF) of a screen film system.

Digital Mammography 87

Bick & Diekmann, 2010). To provide high quality images, all of these steps need to be optimized. As a result, detectors are characterized by their quantum efficiency, sensitivity,

Quantum interaction efficiency describes the quantity of x-rays that reach the detector and interacts with it to produce signal. The quantum interaction efficiency of a detector can be increased by increasing the thickness of the detector. However, quantum interaction efficiency can be reduced by using higher energies since this will decrease the x-ray attenuation coefficient (Pisano et al., 2004). An exception to this is when the x-ray energy exceeds an absorption edge of the detector material, as in CsI. Thus, quantum detection efficiency will influence the sensitivity of detector. Detector sensitivity will also be dependent on the amount of energy required to produce an electron or light quantum to be measured¸ the efficiency of signal collection and measurement of the charge produced

Quantum noise (or mottle) is the result of random fluctuation in the x-ray beam. It is independent of breast density composition. Quantum noise can be statistically described by the Poisson distribution (Pisano et al., 2004). To decrease the quantum noise of an image (increase the signal-to-noise ratio), the amount of x-rays absorbed by the detector have to be increased. This can be performed by using a detector with better quantum detection

Another source of fluctuation or noise that decreases image quality is structural noise. With the use of a detector, digital mammography has eliminated the structural noise of film granularity (random structure of the grains of silver halide) (Bunch, in Van Metter & Beutel, 1997). However, in digital mammography there is some structural difference across the detector and this is associated with spatial variations in detector sensitivity. Because these differences may remain constant over time, they do no represent traditional noise. In digital mammography this is referred to as "fixed pattern noise" or "structural noise". Through the use of image correction with flat-fielding or gain correction this can be removed as seen in

There are two main types of digital mammography imaging systems. One type uses a fullfield detector to be imaged, Figure 7. The detector in this system is stationary and the system may utilize a grid to remove x-ray scatter, thereby increasing the signal-to-noise ratio. The second major type of digital mammography system is a scanned-slot device that uses a detector rectangular in shape, Figure 8. The detector in this type of mammography system scans/moves across the inferior portion of the breast support at the same time a collimated x-ray beam moves during the image acquisition. The detector and the x-ray beam move in synchrony. In this latter type of system, no grid is needed since there is less scatter radiation from a narrower x-ray beam. Regardless of system type, it is important that the system is able to image as close to the chest wall as possible and for the system to accommodate all

spatial resolution, noise, dynamic range and linearity of response.

**3.1 Quantum efficiency & noise** 

(Yaffe, in Bick & Diekmann, 2010).

Figure 6.

breasts sizes.

**3.2 Detector systems** 

efficiency or by increasing exposure (mAs).

Spatial sampling is unique to digital mammography and affects resolution. The signal from each detector element (del) is averaged over the sensitive region or aperture (d). This will result in a decrease of the MTF of a detector (Pisano et al., 2004). The size of the del will supply the information displayed in one pixel. Dels can range from ~ 44 to 100 μm (0.04-0.1 mm). Dels are arranged with a specified center-to-center distance or pitch (p), Figure 5. If the pitch is too large, information will be lost in the sampling process.

Fig. 5. Example of a detector composed of detector elements (d) that contain a sensitive region called the aperture. The distance between the centers of dels is the pitch (p).

#### **2.4 Radiation dose**

The flexibility of digital with decoupling allows for decreased radiation dose compared to SFM. Several factors account for this decreased dose. First, image brightness (display) is now independent of acquisition. Since brightness is not dependent on the amount of x-ray exposure needed to produce the image, it allows the user to determine the dose selection. Secondly, digital detectors have higher detective quantum efficiency with decreased signalto-noise-ratios than a SFM image receptor (Pisano et al., 2004). Consequently, a more penetrating x-ray beam can be used with digital mammography, and this result in a lower patient dose. Currently, some digital mammography systems have dose reductions of 25- 30% compared with SFM (Heddson, 2007; Hendrick, 2010; Yaffe, in Bick & Diekmann, 2010).

## **3. Digital mammography detectors**

The detector is one of the key components of a digital mammography system. It produces an electronic signal that represents the pattern of x-rays transmitted by the breast. Optimally, a detector should include the entire range of x-ray intensities transmitted by different areas of the breast without loss of information. Besides the detector interacting with x-rays transmitted by the breast and absorption of the energy carried by the x-rays, it performs several other important functions. These other functions (in order) include: conversation of the transmitted and absorbed energy to a usable signal (light or electronic charge), collection of this signal, secondary conversion if needed (phosphor-based detectors), readout of the charge, amplification, and finally digitization of the information (Pisano et al., 2004; Yaffe, in

Spatial sampling is unique to digital mammography and affects resolution. The signal from each detector element (del) is averaged over the sensitive region or aperture (d). This will result in a decrease of the MTF of a detector (Pisano et al., 2004). The size of the del will supply the information displayed in one pixel. Dels can range from ~ 44 to 100 μm (0.04-0.1 mm). Dels are arranged with a specified center-to-center distance or pitch (p), Figure 5. If

Fig. 5. Example of a detector composed of detector elements (d) that contain a sensitive region called the aperture. The distance between the centers of dels is the pitch (p).

The flexibility of digital with decoupling allows for decreased radiation dose compared to SFM. Several factors account for this decreased dose. First, image brightness (display) is now independent of acquisition. Since brightness is not dependent on the amount of x-ray exposure needed to produce the image, it allows the user to determine the dose selection. Secondly, digital detectors have higher detective quantum efficiency with decreased signalto-noise-ratios than a SFM image receptor (Pisano et al., 2004). Consequently, a more penetrating x-ray beam can be used with digital mammography, and this result in a lower patient dose. Currently, some digital mammography systems have dose reductions of 25- 30% compared with SFM (Heddson, 2007; Hendrick, 2010; Yaffe, in Bick & Diekmann, 2010).

The detector is one of the key components of a digital mammography system. It produces an electronic signal that represents the pattern of x-rays transmitted by the breast. Optimally, a detector should include the entire range of x-ray intensities transmitted by different areas of the breast without loss of information. Besides the detector interacting with x-rays transmitted by the breast and absorption of the energy carried by the x-rays, it performs several other important functions. These other functions (in order) include: conversation of the transmitted and absorbed energy to a usable signal (light or electronic charge), collection of this signal, secondary conversion if needed (phosphor-based detectors), readout of the charge, amplification, and finally digitization of the information (Pisano et al., 2004; Yaffe, in

**2.4 Radiation dose** 

**3. Digital mammography detectors** 

the pitch is too large, information will be lost in the sampling process.

Bick & Diekmann, 2010). To provide high quality images, all of these steps need to be optimized. As a result, detectors are characterized by their quantum efficiency, sensitivity, spatial resolution, noise, dynamic range and linearity of response.

#### **3.1 Quantum efficiency & noise**

Quantum interaction efficiency describes the quantity of x-rays that reach the detector and interacts with it to produce signal. The quantum interaction efficiency of a detector can be increased by increasing the thickness of the detector. However, quantum interaction efficiency can be reduced by using higher energies since this will decrease the x-ray attenuation coefficient (Pisano et al., 2004). An exception to this is when the x-ray energy exceeds an absorption edge of the detector material, as in CsI. Thus, quantum detection efficiency will influence the sensitivity of detector. Detector sensitivity will also be dependent on the amount of energy required to produce an electron or light quantum to be measured¸ the efficiency of signal collection and measurement of the charge produced (Yaffe, in Bick & Diekmann, 2010).

Quantum noise (or mottle) is the result of random fluctuation in the x-ray beam. It is independent of breast density composition. Quantum noise can be statistically described by the Poisson distribution (Pisano et al., 2004). To decrease the quantum noise of an image (increase the signal-to-noise ratio), the amount of x-rays absorbed by the detector have to be increased. This can be performed by using a detector with better quantum detection efficiency or by increasing exposure (mAs).

Another source of fluctuation or noise that decreases image quality is structural noise. With the use of a detector, digital mammography has eliminated the structural noise of film granularity (random structure of the grains of silver halide) (Bunch, in Van Metter & Beutel, 1997). However, in digital mammography there is some structural difference across the detector and this is associated with spatial variations in detector sensitivity. Because these differences may remain constant over time, they do no represent traditional noise. In digital mammography this is referred to as "fixed pattern noise" or "structural noise". Through the use of image correction with flat-fielding or gain correction this can be removed as seen in Figure 6.

#### **3.2 Detector systems**

There are two main types of digital mammography imaging systems. One type uses a fullfield detector to be imaged, Figure 7. The detector in this system is stationary and the system may utilize a grid to remove x-ray scatter, thereby increasing the signal-to-noise ratio. The second major type of digital mammography system is a scanned-slot device that uses a detector rectangular in shape, Figure 8. The detector in this type of mammography system scans/moves across the inferior portion of the breast support at the same time a collimated x-ray beam moves during the image acquisition. The detector and the x-ray beam move in synchrony. In this latter type of system, no grid is needed since there is less scatter radiation from a narrower x-ray beam. Regardless of system type, it is important that the system is able to image as close to the chest wall as possible and for the system to accommodate all breasts sizes.

Digital Mammography 89

Fig. 8. Example of a scanned-slot digital mammography system. a) Pictured is a Fisher

Many different types of detectors are used for digital mammography and these will be briefly described in this section. The first four discussed are used in direct radiography (DR) mammography systems and the last one is used in a computed radiography (CR) system.

A phosphor flat panel detector, Figure 9, is constructed of a plate of amorphous silicon. Through solid-state manufacturing, a rectangular array of light-sensitive photodiodes with a layer of thallium-activated cesium iodide phosphor, CsI (Tl) are deposited onto the plate. The photodiodes are the dels of the detector. These dels will detect the light emitted by the

Besides each del containing a photodiode, it also contains a thin film transistor (TFT) switch, and these are interconnected with an array of control and data lines. A readout line is present along each column of the detector, and when a control line is activated it activates all the TFTs in that row (Yaffe, in Bick & Diekmann, 2010). The signal from the row of activated dels is then transferred to an amplifier and digitizer. The digitized information

from one del will represent the information corresponding to a pixel of the image.

SenoScan mammography unit. b) Image of the detector.

a) b)

phosphor and create and store an electric charge.

**3.3 Detector types** 

**3.3.1 Phosphor flat panel** 

Fig. 6. Example of flat-field correction. a) Uncorrected image. b) Image after flat-field correction. Courtesy of Martin Yaffe, PhD; Sunnybrook Health Sciences Centre, Toronto, Canada. (from Digital Mammography, eds. ED Pisano, MJ Yaffe, CM Kuzmiak. Lippincott, Williams & Wilkins, a Walters Kluwer Company, 2004. With permission.)

Fig. 7. Example of a FFDM detector system. Pictured is the General Electric Senographe Essential.

Fig. 6. Example of flat-field correction. a) Uncorrected image. b) Image after flat-field correction. Courtesy of Martin Yaffe, PhD; Sunnybrook Health Sciences Centre, Toronto, Canada. (from Digital Mammography, eds. ED Pisano, MJ Yaffe, CM Kuzmiak. Lippincott,

Fig. 7. Example of a FFDM detector system. Pictured is the General Electric Senographe

Essential.

Williams & Wilkins, a Walters Kluwer Company, 2004. With permission.)

Fig. 8. Example of a scanned-slot digital mammography system. a) Pictured is a Fisher SenoScan mammography unit. b) Image of the detector.

#### **3.3 Detector types**

Many different types of detectors are used for digital mammography and these will be briefly described in this section. The first four discussed are used in direct radiography (DR) mammography systems and the last one is used in a computed radiography (CR) system.

#### **3.3.1 Phosphor flat panel**

A phosphor flat panel detector, Figure 9, is constructed of a plate of amorphous silicon. Through solid-state manufacturing, a rectangular array of light-sensitive photodiodes with a layer of thallium-activated cesium iodide phosphor, CsI (Tl) are deposited onto the plate. The photodiodes are the dels of the detector. These dels will detect the light emitted by the phosphor and create and store an electric charge.

Besides each del containing a photodiode, it also contains a thin film transistor (TFT) switch, and these are interconnected with an array of control and data lines. A readout line is present along each column of the detector, and when a control line is activated it activates all the TFTs in that row (Yaffe, in Bick & Diekmann, 2010). The signal from the row of activated dels is then transferred to an amplifier and digitizer. The digitized information from one del will represent the information corresponding to a pixel of the image.

Digital Mammography 91

to increased line-spread function in SFM, CsI crystals used in digital systems are more efficient at transferring the light produced (Pisano et al., 2004). This increase in efficiency is because the CsI crystals act as fiber optics. Consequently, the detector can be made thicker

An example of a commercial system with this detector is produced by General Electric Medical Systems (Milwaukee, WI), Figure 7. The field size is 24 cm x 31 cm, the del pitch is 100 μm, and the digitization is 14 bits (Ghetti et al., 2008). Of interest, to correct for inhomogeneous areas in the detector, flat-fielding or gain correction requires that an offset

A phosphor-CCD system also uses CsI(Tl) as the material for x-ray absorption to light conversion in the detector. However, the CsI(Tl) is deposited on a rectangular fiber-optic coupling plate. The fibers conduct the light from the CsI to a charge-coupled device (CCD) array. The CCD is an electronic chip containing rows and columns of light-sensitive

The phosphor-CCD system detector is long, narrow and rectangular in shape, approximately 1 cm x 24 cm as seen in Figure 8b. The x-ray beam is collimated into a narrow band since it and the detector scan across the breast in synchrony, Figure 10. The charge created in the CCD is transferred down the columns from row to row at the same rate, but in opposite direction to the physical motion of the detector. The bundles of charges are integrated, collected and read out corresponding to x-ray transmission on the detector for each x-ray path through the breast (Pisano et al., 2004). This is known as time-delay

elements. The CCD converts the light into an electronic signal that is digitized.

Fig. 10. Schematic of the path of detector travel in a scanned-slot detector system.

value and a gain be measured for each del (Pisano et al., 2004).

without loss of resolution.

**3.3.2 Phosphor-CCD system** 

integration.

#### b)

Fig. 9. Illustrations of a CsI-amorphous silicon photodiode flat panel detector. a) Generic detector. Courtesy of Martin Yaffe, PhD; Sunnybrook Health Sciences Centre, Toronto, Canada (from Digital Mammography, eds. ED Pisano, MJ Yaffe, CM Kuzmiak. Lippincott, Williams & Wilkins, a Walters Kluwer Company, 2004. With permission. b) Commercial detector. (Courtesy of General Electric Medical Systems, Milwaukee, WI).

In this type of detector, CsI is used because of its crystal structure. These crystals can be commercially grown to form needle-like or columnar structures. Unlike granular phosphors that allow the produced light upon x-ray absorption to move laterally in the system leading

Fig. 9. Illustrations of a CsI-amorphous silicon photodiode flat panel detector. a) Generic detector. Courtesy of Martin Yaffe, PhD; Sunnybrook Health Sciences Centre, Toronto, Canada (from Digital Mammography, eds. ED Pisano, MJ Yaffe, CM Kuzmiak. Lippincott, Williams & Wilkins, a Walters Kluwer Company, 2004. With permission. b) Commercial

In this type of detector, CsI is used because of its crystal structure. These crystals can be commercially grown to form needle-like or columnar structures. Unlike granular phosphors that allow the produced light upon x-ray absorption to move laterally in the system leading

detector. (Courtesy of General Electric Medical Systems, Milwaukee, WI).

a)

b)

to increased line-spread function in SFM, CsI crystals used in digital systems are more efficient at transferring the light produced (Pisano et al., 2004). This increase in efficiency is because the CsI crystals act as fiber optics. Consequently, the detector can be made thicker without loss of resolution.

An example of a commercial system with this detector is produced by General Electric Medical Systems (Milwaukee, WI), Figure 7. The field size is 24 cm x 31 cm, the del pitch is 100 μm, and the digitization is 14 bits (Ghetti et al., 2008). Of interest, to correct for inhomogeneous areas in the detector, flat-fielding or gain correction requires that an offset value and a gain be measured for each del (Pisano et al., 2004).

### **3.3.2 Phosphor-CCD system**

A phosphor-CCD system also uses CsI(Tl) as the material for x-ray absorption to light conversion in the detector. However, the CsI(Tl) is deposited on a rectangular fiber-optic coupling plate. The fibers conduct the light from the CsI to a charge-coupled device (CCD) array. The CCD is an electronic chip containing rows and columns of light-sensitive elements. The CCD converts the light into an electronic signal that is digitized.

The phosphor-CCD system detector is long, narrow and rectangular in shape, approximately 1 cm x 24 cm as seen in Figure 8b. The x-ray beam is collimated into a narrow band since it and the detector scan across the breast in synchrony, Figure 10. The charge created in the CCD is transferred down the columns from row to row at the same rate, but in opposite direction to the physical motion of the detector. The bundles of charges are integrated, collected and read out corresponding to x-ray transmission on the detector for each x-ray path through the breast (Pisano et al., 2004). This is known as time-delay integration.

Fig. 10. Schematic of the path of detector travel in a scanned-slot detector system.

Digital Mammography 93

Fig. 11. An illustration of the Fuji FFDM system that uses two layers of selenium. (Courtesy

system has a del size of 50 μm and a bit depth of 16. In the system by XCounter (Stockholm, Sweden), it uses a set of multiple linear detectors and scans across the breast in synchrony with a collimated x-ray beam similar to the Sectra system. However, it uses a pressurized gas as the x-ray absorber. The pulses of ions generated by the gas form the signal (Thunberg, in Antonuk & Yaffe, 2002). The XCounter system has the same del size and bit depth as the

The last detector to be discussed is the PSP system which is a computed radiography (CR) system. CR systems have been in use in general radiography for many years and are based on the principle of photostimulable luminescence. More recently they have been developed and used in mammography. The CR mammography systems utilize a phosphor screen. Energy from x-ray absorption causes electrons in the phosphor crystal to be liberated from the matrix and captured and stored in "traps" in the crystal lattice (Pisano et al., 2004), as seen in Figure 13. The number of traps filled is proportional to the amount of absorbed x-ray

of Fujifilm Medical, Stamford, CT).

Sectra. Neither system uses a grid.

signal.

**3.3.5 Photostimulable Phosphor (PSP) System** 

A major advantage of this scanned-slot system is the result of the x-ray beam being collimated and only part of the breast being imaged at a time. Consequently, the transmitted x-rays are not lost (scatter-to-primary ratio is reduced), resulting in a grid no longer being needed. Therefore, the dose should be reduced. A limitation of this type of system is that it requires a longer image acquisition time.

A commercial unit like this was originally marketed by Fischer Imaging Inc (Denver, CO) as seen in Figure 8. It has dels of 54 μm with digitization performed at 12 bits. Of interest, over a small area of the detector, data could be read out at 27 μm to provide a high-resolution mode.
