A comparison of digital mammography detectors and emerging technology

Radiography xxx (2015) 1e9

Contents lists available at ScienceDirect

Radiography journal homepage: www.elsevier.com/locate/radi

A comparison of digital mammography detectors and emerging technology J.L. Diffey* Hunter New England Imaging, John Hunter Hospital, Lookout Road, New Lambton Heights, NSW 2305, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 April 2015 Received in revised form 25 June 2015 Accepted 29 June 2015 Available online xxx

The overall diagnostic accuracy of digital mammography in the context of screening has been shown to be similar or slightly better than screen-film mammography. However, digital mammography encompasses both Computed Radiography (CR) and integrated Digital Radiography (DR) and there is increasing evidence to suggest that differences in detector technology are associated with variations in cancer detection rate, dose and image quality. These differences are examined in detail. Although digital mammography offers many advantages compared to screen-film, there are still some limitations with its use as a screening tool and reduced cancer detection in dense breasts remains an issue. Digital mammography detectors have paved the way for emerging technologies which may offer improvements. Taking the definition of mammography to only include X-ray imaging of the breast, this article focuses on tomosynthesis, contrast-enhanced digital mammography, stereoscopic mammography and dedicated breast computed tomography. Advanced software applications such as Computed Aided Detection (CAD) and quantitative breast density assessment are also presented. The benefits and limitations of each technique are discussed. © 2015 The College of Radiographers. Published by Elsevier Ltd. All rights reserved.

Keywords: Breast cancer Breast density Breast tomosynthesis Contrast enhanced Digital mammography

Introduction Breast cancer is a major health burden on a global scale, with over 1.6 million new cases being diagnosed worldwide each year.1 As the population ages, risk and consequently incidence are expected to rise, making breast screening as important now as when it was introduced. A national breast screening programme commenced in Australia in 19912 and screening is now carried out exclusively using digital mammography. The term digital mammography (DM) encompasses both Computed Radiography (CR) and integrated Digital Radiography (DR), with DR incorporating a range of detector technologies, shown in Fig. 1. There are a number of advantages to DM compared to screenfilm. Film was both the detector and the display; in DM these are separate devices, enabling each stage of imaging (acquisition, preprocessing, post-processing and display), to be optimised independently of the others.

* Hunter New England Imaging, Department of Nuclear Medicine, John Hunter Hospital, Lookout Road, New Lambton Heights, NSW 2305, Australia. Tel.: þ61 (0)2 4921 3380. E-mail address: [email protected].

The major strength of digital detectors is their wide dynamic range. The relationship between grey level and exposure is linear, as opposed to film which follows an S-shaped characteristic curve. Image contrast is generated when grey level changes with exposure. For film, this only occurs over a narrow range of exposures; for digital detectors the exposure range, or dynamic range, is much wider. Optimisation is therefore essential because the correct exposure is no longer limited purely by contrast, but also by noise. If the dose is too low, the image will have unacceptably high quantum mottle. If the dose is too high, the patient will receive unnecessary radiation dose, which could go unnoticed because unlike film, digital image processing prevents image saturation. This phenomenon is known as dose creep in digital imaging. Fortunately, the stringent quality control in mammography and the introduction of a new parameter, known as Signal Difference to Noise Ratio (SDNR) to achieve the optimum balance between dose and image quality,3 means that dose creep has not been an issue. Several clinical dose audits have shown that patient doses are actually lower for DR systems than for film.4e7 This is partially attributed to the use of harder beam qualities and most digital systems now employ a tungsten target instead of, or in addition to, molybdenum and rhodium. A heavily filtered tungsten spectrum (to remove the undesirable L-characteristic radiation) results in more efficient X-ray

http://dx.doi.org/10.1016/j.radi.2015.06.007 1078-8174/© 2015 The College of Radiographers. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

2

J.L. Diffey / Radiography xxx (2015) 1e9

Figure 1. Detector technologies employed in analogue and digital mammography.

production and a higher effective energy.8,9 The loss in physical contrast associated with higher energy X-rays can be compensated for by increasing the detector dose, which reduces quantum noise in the image.3 DM has greater contrast resolution than film, which in principle should lead to improved diagnostic accuracy, particularly in women with dense breasts. This has been found to be true in practice; studies have shown that the overall diagnostic accuracy of DM and screen-film mammography in the context of screening is comparable,10e13 or slightly better.14,15 Analysis of particular subgroups found that DM was more sensitive in women with radiologically dense breasts and in younger women.10,14,16 Image processing may also be credited with improved sensitivity in dense breasts, as radiologists have reported that breasts appear to be less dense on digital mammography images compared to film.17 The limiting spatial resolution of DM (5e10 lp/mm) is lower than that of screen-film (15e20 lp/mm). However, despite initial concerns that this may lower microcalcification detection rates, the opposite has been found for DR detectors, which exhibit significantly greater sensitivity than screen-film.14,16,18e22,26,27 Neal et al. found that the detection rate of high-risk lesions, which often manifest as calcifications, was three times higher for DM compared to film.19 Although this might facilitate risk-reduction strategies, it may contribute to over-surveillance and over-treatment, both of which are common criticisms of mammographic screening.23 Practical advantages of digital mammography include easier archival, retrieval and transmission of electronic images and potentially higher patient workflow for DR, but possibly not for CR. Comparison of digital mammography detectors A CR image plate is a layer of photostimulable phosphor on a transparent plastic substrate, contained within a cassette which slots into the Bucky in the same manner as a screen-film cassette. CR can therefore be used with existing mammography X-ray units and is considered a cost-effective solution into DM. CR plates must be read in order to view the image, but are reusable. The effective detector element (del) size is typically 50 mm and spatial resolution of 10 lp/mm is theoretically possible.24 With indirect conversion detectors, X-rays are first converted to light photons in a scintillation layer (usually Caesium Iodide). Light

photons are then converted to an electronic charge signal using a flat panel of amorphous silicon (a-Si) that incorporates an array of photodiodes. The effective del size is typically 100 mm, corresponding to a maximum achievable spatial resolution of 5 lp/mm.24 Direct conversion flat panel detectors utilise a semiconductor material known as amorphous selenium (a-Se) to convert X-rays directly to electronic charge. Nominal del sizes are 50e85 mm, which equates to a limiting spatial resolution of 6e10 lp/mm. This is the technology employed by most digital mammography vendors. An alternative method developed by one vendor employs single x-ray photon counting with energy discrimination thresholds to reject scattered photons and electronic noise. X-ray photons are then converted directly to electronic signal in a crystal silicon detector.24 Since the publication of the Digital Mammography Imaging Screening Trial (DMIST)10 ten years ago, research has shifted from a comparison of DM with screen-film, to the variation in performance of different digital detector technologies. Recently published studies have compared cancer detection rates,16,21,22,25e27 patient dose4e7,28 and image quality.7,27,29,30 Keavey et al. reported no significant difference in overall cancer detection rates between digital mammography systems employing three different detector technologies.24 However, all of these were DR detectors. Chiarelli et al. found that overall cancer detection rates were lower with CR (3.4 per 1000 women screened) compared to film (4.8 per 1000) or DR (4.9 per 1000) with CR significantly less likely to help detect invasive cancers.26 Although a similar trend in overall cancer detection rates was reported by Seradour et al. (CR: 5.5 per 1,000, film: 6.6 per 1000 and DR: 7.1 per 1000), there was no significant difference in invasive cancer detection.16 Investigations in Australia also showed no significant difference in invasive and small invasive cancer detection rates between CR and DR; these studies also concluded that CR was at least as good as film, based on overall cancer detection rates.21,22 The only common finding between all studies was that DR was significantly better at detecting ductal carcinoma in situ (DCIS).16,21,22,26 All of these studies16,21,22,26 examined concurrent cohorts of a screening population (aged 50e74) and all mammograms were double-read, except for those in the study by Chiarelli et al.26 which were single-read; however, this would have been the case for film, CR and DR images. Differences in the study results may be due to differences in the CR systems and it would be interesting to see not just a comparison of CR and DR, but a comparison by system manufacturer. 79% of the CR systems used in the study by Chiarelli et al.26 were from a vendor, which initially failed to comply with European image quality standards31 and was deemed unsuitable for use in the UK breast screening programme,32 although a later model by the same vendor was considered acceptable following optimal adjustment of the Automatic Exposure Control.33 The lower cancer detection rates for CR compared to DR, particularly for DCIS, are somewhat expected, based on assessment of physical image quality measures.7,27,29,30 The parameters of interest are spatial resolution, contrast and noise (which has quantum, structural and electronic components).8,29,34 Each parameter exhibits a dependence on the others, so overall performance is commonly evaluated using metrics such as the modulation transfer function (MTF), signal to noise ratio (SNR) and detective quantum efficiency (DQE)8,27,29 Yaffe et al. found that, despite their smaller nominal del size, the MTF was actually lower for CR than DR systems.29 Explanations for this include a degradation in spatial resolution caused by scattering of laser light in the phosphor layer and inefficient conversion of X-ray photons to electronic signal within the photostimulable phosphor and readout processes.29 The lower DQE of CR compared to DR is associated with a lower SNR, which is

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

J.L. Diffey / Radiography xxx (2015) 1e9

reflected in poorer visibility of certain anatomic structures, particularly calcifications.27 CR therefore requires more X-ray photons to meet image quality requirements and clinical mean glandular doses (MGD) are higher for CR than DR.4e7 Fig. 2 shows the MGD required to achieve the same level of image quality, assessed using the CDMAM threshold contrast detail detectability phantom (Artinis Medical Systems).7 It can be seen that although all technologies attain acceptable image quality at a dose below the remedial level, the MGD required is lower than film for DR and higher than film for all but one CR system. The data presented in Fig. 2 is for new systems. The author has observed a decrease in the sensitivity of CR plates over time, necessitating increased dose to meet Australian image quality requirements.35,36 Measurements from medical physics quality control testing at breast screening sites within Australia are shown in Table 1. The average increase in dose required by CR systems (image plates and readers) greater than 3 years old was 18%. In light of the above, the Australasian College of Physical Scientists and Engineers in Medicine (ACPSEM) recommends that “only DR technology should be approved for future purchases of equipment for screening mammography in Australia and New Zealand and existing CR systems should be progressively replaced.”24 Emerging technologies The aim of breast screening is to reduce mortality from breast cancer whilst minimising physical and psychological harm; overdiagnosis remains a common criticism.23 Despite the advantages of DM compared to screen-film imaging, there are still some limitations. The ideal system would detect more invasive cancer with greater confidence, improve diagnostic accuracy in dense breasts and reduce recall rates. DR detectors have paved the way for emerging technologies such as digital breast tomosynthesis, contrast enhanced digital mammography and stereoscopic mammography, which aim to fulfil these requirements. Ultrasound, magnetic resonance imaging (MRI) and nuclear medicine techniques are not discussed, but their role as beneficial adjunct techniques in the field of breast cancer diagnosis is acknowledged. Digital breast tomosynthesis (DBT) DBT involves the acquisition of a series of low-dose projection images in an arc around the compressed breast, shown schematically in Fig. 3.

Figure 2. Mean glandular doses required to reach acceptable image quality for 0.1 mm detail of CDMAM phantom; Figure shows data from NHSBSP Report 1304.7

3

The individual projection images are reconstructed into slices through the breast, typically 1 mm thick with planes parallel to the breast support platform.37e43 Image reconstruction enhances objects of interest within each slice and blurs out background objects from other planes. Because several projection images are required, each must be acquired at a much lower dose than a conventional DM image in order to achieve a similar examination dose. DBT projection images therefore have more quantum noise and it is essential that the digital detectors utilised have excellent X-ray absorption efficiency (a high DQE).38 DBT is referred to as three dimensional (3-D) mammography, although the term quasi 3-D is strictly more appropriate since the projection images are acquired over a limited angular range, typically 15e25 ,41e45 although some vendors use 40e50 .41,43,45 The advantages of a wide angle are better depth resolution, reduced tissue overlap43,46 and an improvement in the appearance of large, low contrast lesions.43 This is balanced against the detriments associated with a longer scan time, such as patient discomfort and motion blur.46 A larger scan angle is potentially associated with more projection images. This enables reconstruction with fewer artefacts, but with increased raw data size and reconstruction times and a compromise must be met between greater total examination dose and lower dose per image (with higher image noise). The number of projections can be reduced by increasing the angular increment, but too few will result in view aliasing artefacts.8,43 The strengths of DBT compared to DM are that it removes superimposition of non-adjacent breast tissues. This enables visualisation of lesions that would be masked by over- or under-lying tissue and distinguishes real lesions from those mimicked by superimposition of normal structures.37e46 These benefits have been realised clinically in a number of studies on large screening populations.47e51 Table 2 shows that all studies found an increase in overall cancer detection rate and a reduction in recall rate when using DBT in addition to DM, compared to using DM alone.37e46 Better clarity of lesions and increased confidence in reading were also reported. More invasive cancers were detected,48e50 with no increase in the detection of in-situ cancer,50,51 which is promising, but studies of duration greater than 10 years are needed to assess the true impact on overdiagnosis.49 The addition of DBT to DM was also found to improve the sensitivity in younger women and those with dense breasts.51 Although earlier studies investigated the use of DBT as an assessment tool,52e54 recent evidence suggests that it may be beneficial as a screening tool,47e51 especially if radiation dose implications are addressed.55 Using DBT in addition to DM is associated with a dose increase of approximately 2e3 times compared to DM alone. Table 3 shows average MGD for each study population, corresponding to a compressed breast thickness of 50e55 mm. The dose for DBT is higher than that for DM, albeit only slightly for the average breast.49,57,58 However, the relative increase is dependent upon breast glandularity, and is observed to be greatest for fatty breasts.58e60 In a study using water- and oil-based phantoms to simulate a range of breast thicknesses and compositions, Feng et al. found that DBT dose was only 8% higher than DM dose for a phantom of thickness 50 mm with 50% glandularity, but 83% higher for a phantom of 60 mm thickness with14% glandularity.58 In a clinical study, Tromans et al. observed that MGD for DBT was generally higher than that of DM for fatty breasts, but was in fact lower than DM for dense breasts.60,61 This finding is somewhat encouraging, based on the fact that women with dense breasts are likely to benefit the most from DBT,51 but raises the question as to whether women with fatty breasts should have DBT in addition to DM, given that there may be fewer benefits from reduced superimposition. However, Skaane et al. reported increases in cancer detection for all breast densities.49

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

4

J.L. Diffey / Radiography xxx (2015) 1e9

Table 1 MGD to the American College of Radiology Mammographic Accreditation Phantom to meet Australian image quality requirements.35,36

CR: new system CR: >3 years old DR Upper limit

MGD to meet image quality requirements (mGy)

Additional information

1.1e1.5 1.2e1.9 0.7e1.3 Remedial: 2 mGy Suspension: 3 mGy

Data from 16 units Data from 50 units, including direct and indirect DR

Figure 3. Principle of Digital Breast Tomosynthesis. The tube rotates in an arc around the compressed breast from position A to C; overlying structures are separated on the projection images.

In order to reduce examination dose, it was hypothesised that one-view DBT would be a sufficient addition to DM, using the medio-lateral oblique view. Several studies have shown that the increased cancer detection with two-view DBT justifies the higher dose compared to one-view DBT,61e63 with lesions being seen on one view only in up to 9% of cases, or seen much more clearly on the

cranio-caudal view in up to 34% of cases.62 However, these were all conducted using a system with an angular range of 15 . Interestingly, in a recent study using a wide-angle tomosynthesis system (50 ), Lang et al. observed significantly higher cancer detection rates with one-view DBT compared to two-view DM and concluded that breast screening may even be feasible using one-view DBT as a stand-alone modality.64 This is an area that warrants further investigation, as recall rates were 43% higher for DBT than DM.64 Instead of acquiring additional DM images, a better solution may be to use a synthesised 2-D image from the DBT data. Skaane et al. reported no significant difference in cancer detection rate between the DM and synthesised image when used in conjunction with DBT in a screening population (7.8 compared to 7.7 per 1000); DM alone had a detection rate of 6.1 per 1000.55 Zuley et al. found that the synthesised image offered improvements in sensitivity and specificity over the DM image, when both were used alone and in conjunction with DBT.65 An additional interesting finding was that when a biopsy proven cancer was present, the radiologists assigned a higher BI-RADS score66 to the synthesised images than to the DM images but this did not lead to a significant difference in recall rates.65 Fig. 4 shows that although synthesised images can look very similar to DM images,67 there may be cases where certain features are emphasised, such as the spiculations on the mass.65 Skaane et al. reported a doubling in reading time for DM þ DBT, compared to DM alone.49 This is somewhat inevitable as an average-sized breast of 55 mm thickness will have 55 slices! However, this is likely to decrease with experience.

Contrast enhanced digital mammography (CEDM) Contrast-enhanced digital (or spectral) mammography (CEDM), utilises the principle of tumour angiogenesis. Both morphology and

Table 2 Results from clinical studies comparing the performance of Digital Mammography (DM) alone to Digital Mammography þ Digital Breast Tomosynthesis (DM þ DBT) in large screening populations.47e51 Study

Population

DM/DM þ DBT

Cancer detection rate per 1000

Recall rate per 1000

Ciatto et al., 201347

7292

DM DM DM DM DM DM DM DM DM DM

5.3 8.1 3.5 5.4 6.1 8.0 4.2 5.4 4.6 5.5

44 35 82 54 67 61 107 91 104 88

Rose et al., 201448

10,878

Skaane et al., 201349

12,621

Friedewald et al., 201450

281,187 173,663 10,728 15,571

McCarthy et al., 201451

þ DBT þ DBT þ DBT þ DBT þ DBT

Table 3 Mean glandular dose (MGD) for digital mammography (DM) and digital breast tomosynthesis (DBT).49,56e58 Study

Population

DM MGD (mGy)

DBT MGD (mGy)

DM þ DMT MGD (mGy)

Increase relative to DM

Skaane et al.49 Cavagnetto et al.56 Olgar et al.57 Feng et al.58

24,901 300 641 Phantoms: 50 mm thickness 50% glandularity

1.58 1.31 1.82 1.20

1.95 2.56 2.53 1.30

3.53 3.87 4.35 2.50

2.24 2.95 2.39 2.08

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

J.L. Diffey / Radiography xxx (2015) 1e9

5

Figure 5. Imaging procedure of dual-energy contrast-enhanced digital mammography; reproduced from Dromain et al.69 with permission from Elsevier.

Figure 4. DM images (left) and synthesised images (right), using Hologic C-View. (a) The synthesised image has a very similar appearance to the DM image. Reproduced from Smith67 with permission from Hologic. (b) Certain features can be emphasised in the synthesised image, in this case the spiculations on the mass, which proved to be invasive ductal carcinoma at biopsy. Reproduced from Zuley et al.65 with permission from the Radiological Society of North America (RSNA®)

enhancement information are used to improve detection and characterisation of breast carcinoma. Two CEDM techniques have been developed: temporal and dual-energy.68,69 During temporal CEDM the breast is compressed lightly to avoid reducing blood flow. A single high-energy image (the mask) is acquired, prior to the injection of iodinated contrast agent. Multiple high-energy images are then acquired. The mask is subtracted from these post-contrast images to enhance contrast uptake areas and uptake-washout curves to assess tumour kinetics can be derived.68,69 Unfortunately, this technique is far less sensitive in CEDM than MRI, with curves often showing gradual uptake for both malignant and benign tumours. Because CEDM is 2D projection imaging, the region of interest is drawn over a column, rather than a slice, of tissue and normal tissue could mask the washout of lesions.68,69 A further disadvantage of this technique is the long examination time (15e20 min), during which the patient remains under compression. Patient discomfort may mean that only a cranio-caudal view is possible and motion and misregistration artefacts are likely.68,69 Dual-energy CEDM, shown in Fig. 5, is therefore the technique of choice.68e71 Iodinated contrast agent is injected into the arm. The

breast is compressed and low and high energy images are acquired in quick succession; the energies are above and below the K-absorption edge of iodine (33.2 keV).8 It is possible to perform two views of both breasts with a single administration of contrast. Images show suppression of glandular tissue and enhancement of contrast uptake; a clinical example is shown in Fig. 6.72 CEDM has demonstrated better sensitivity and specificity than DM,72e74 particularly in dense breasts.73,74 Sensitivity has been found to be equivalent to MRI but with better specificity.75,76 Furthermore, CEDM is faster and cheaper than MRI and images share the appearance of DM. It could be performed immediately after mammography in the case of questionable findings (for example in “one-stop” clinics), thereby reducing the anxiety of being recalled for further investigations. However, it is recommended that high-risk women are examined using MRI due to their increased radiosensitivity.76 A significant limiting factor for CEDM is the requirement for contrast agent, which may be contraindicated in some women and therefore prevents its use as a screening tool. Another disadvantage is the increased radiation dose if it is to be used in addition to DM. However, this may not be necessary in the assessment clinic as it has been shown that the low-energy mammogram obtained during CEDM can be used in lieu of the DM image.73,74 Furthermore, CEDM may be an alternative to spot and magnification views. CEDM is currently suitable for diagnostic imaging only. However, a technique known as non-contrast single-shot spectral mammography has been developed77,78 which could in theory be used for screening. It uses photon counting technology with additional energy thresholds.77 A single X-ray exposure provides the conventional mammogram image, an energy-subtracted image to visualise tumours and breast density quantification, with no additional radiation dose.78 Stereoscopic mammography The principles of stereoscopic mammography are shown in Fig. 7.79 Two DM images are acquired with a 10 angular difference and each is displayed on a high resolution monitor. Monitors are separated by 110 and bisected by a glass plate with a half-silvered coating. This means that one image is transmitted, whilst the other is reflected. By wearing cross-polarised 3-D glasses so that one eye sees only the transmitted image and the other sees only the reflected image, the radiologist's visual system fuses the two images into a single in-depth image of the internal structure of the breast.80

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

6

J.L. Diffey / Radiography xxx (2015) 1e9

lower noise than that in the individual mammography images,82 dose reduction is theoretically possible and is currently about 1.3 times the dose of DM.79 Other limitations associated with the technique are that mammograms are presented horizontally, not vertically, and radiologists need to have functional depth perception. These factors, combined with the need for new, dedicated workstations have so far prevented its widespread use. However, given that DBT images are acquired at a series of angles around the breast, it may in fact be possible to create stereo pairs from DBT images, if image noise is acceptable; this requires further investigation. Dedicated breast computed tomography (DBCT) Conventional CT is unsuitable for mammography as projections through the entire thorax would contribute to unacceptable patient radiation dose. In DBCT, the patient lies prone and the breast hangs through an opening in the bed, as shown in Fig. 8; no compression is required. A cone-beam of X-rays and a flat panel detector make a single rotation around the breast in less than 10 s83e85 DBCT images are not compromised by tissue superimposition.84e87 It therefore shows improved sensitivity in dense breasts86 and volumetric breast density quantification is possible.84 The radiation dose is similar to that from a two-view screening mammography examination, typically 2e4 mGy.84,85 It is not yet intended for breast screening, but one vendor recently received FDA (US Food and Drug Administration) approval for use in breast cancer diagnosis.83,84 DBCT may be used in the assessment of women, where previously MRI or ultrasound were indicated. Although these have the advantage of using no ionising radiation, DBCT has higher spatial resolution, quicker examination times and images have a similar appearance to mammograms.86,87 Figure 6. Typical example of contrast enhanced digital mammography (CEDM); top row shows low energy images, bottom row shows subtracted images (demonstrating enhancement of lesions). The arrow indicates an irregular density in the right craniocaudal view. CEDM demonstrated the true extent of disease and also the multifocal nature, which was confirmed by histopathological analysis of the surgical specimen. Reproduced from Lobbes et al.72 with permission from Springer.

Studies performed on a population of high-risk women, screened using both standard and stereoscopic mammography showed significant decreases in recall rate and improvements in specificity, with a slight but insignificant increase in sensitivity.80,81 The radiation dose associated with stereoscopic mammography in these studies was double that of standard DM.80,81 Stereoscopic mammography could potentially be a screening tool, if radiation dose was reduced. Because the fused image has

Advanced software applications Even without upgrading to new imaging technologies, advanced software applications may provide additional useful information from existing mammography images. Breast density assessment Breast density is a well-established risk factor for breast cancer and is negatively associated with the diagnostic sensitivity of mammography.88e96 Breast density notification laws have now been passed in 21 States of the USA96 since knowledge of breast density is used to inform follow-up imaging, such as ultrasound or MRI. There is interest in incorporating breast density into risk prediction models,97e101 which could be used to tailor screening intervals to individual risk.102e104

Figure 7. Principles of stereoscopic mammography.79 Reproduced with permission from FUJIFILM Australia Pty Ltd.

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

J.L. Diffey / Radiography xxx (2015) 1e9

7

References

Figure 8. Koning Corporation Dedicated Breast CT system. Image courtesy of US Food and Drug Administration website.83

Although numerous techniques have been developed for breast density estimation,105e112 often showing excellent correlation with breast cancer risk, some are subjective105,106 and time-consuming.106e108 If breast density measurement is to be routinely used, the method must be objective and fully-automated. Volumetric techniques fulfilling these requirements now exist, such as Quantra111 and Volpara™,112 In a recent case-control study113 evaluating these two volumetric techniques111,112 and four areabased techniques, Eng et al. found that all methods demonstrated a strong, positive association between percentage breast density (PD) and breast cancer risk, with the increase in risk per increment PD being highest for Volpara™.113

Computer aided detection (CAD) In many countries, it is standard practice for screening mammograms to be independently read by two radiologists. CAD has been investigated as an alternative to the second reader for economic or litigation reasons, or due to difficulties recruiting radiologists. Potential abnormalities on the image are highlighted and an early concern was high false positive rates.114 In a large prospective study comparing double reading to single reading plus CAD, no significant difference in sensitivity was observed although there was a small but significant increase in recall rate for single reading plus CAD.115 In a recent retrospective study, 300 examinations from DMIST10 were re-read by 15 radiologists using two CAD systems.116 No significant difference in sensitivity and specificity was found for the group as a whole.116 CAD is currently being investigated for DBT and may potentially reduce reading times.117

Conclusion Digital mammography offers a number of advantages compared to screen-film mammography and advances in digital detectors have paved the way for exciting new technologies. These have shown great promise in increasing the detection of invasive breast cancer, improving the diagnostic accuracy in dense breasts and reducing recall rates.

Conflict of interest

1. World Health Organisation International Agency for Research on Cancer. GLOBOCAN 2012: estimated cancer incidence, mortality and prevalence worldwide in 2012. Available from: http://globocan.iarc.fr/Pages/fact_sheets_ cancer.aspx [accessed 17.03.15]. 2. BreastScreen Australia. Australian Government Department of Health. Available from: http://www.cancerscreening.gov.au/internet/screening/publishing. nsf/Content/breastscreen-n-you-html [accessed 17.03.15]. 3. Young KC, Oduko JM, Bosmans H, Nijs K, Martinez L. Optimal beam quality selection in digital mammography. Br J Radiol 2006;79:981e90. 4. Oduko JM, Young KC, Burch A. A survey of patient doses from digital mammography systems in the UK in 2007 to 2009. In: Marti J, Oliver A, Freixenet J, Marti R, editors. Proceedings of the 10th International Workshop on digital mammography. Berlin: Springer-Verlag; 2010. p. 365e70. LNCS 6136. 5. Hauge IHR, Pedersen K, Sanderud A, Hofvind S, Olerud HM. Patient doses from screen-film and full-field digital mammography in a population-based screening programme. Radiat Prot Dosim 2012;148(1):65e73. 6. Hendrick RE, Pisano ED, Averbukh A, Moran C, Berns EA, Yaffe MJ. Comparison of acquisition parameters and breast dose in digital mammography and screen-film mammography in the American College of Radiology Imaging Network Digital Mammographic Imaging Screening Trial. AJR Am J Roentgenol 2010;194:362e9. 7. Oduko JM, Young KC, Warren L. Technical evaluation of the Fuji Amulet f/s digital breast imaging system, NHSBSP report 1304. NHS Cancer screening programmes. 2013. 8. Bushberg JT, Seibert JA, Leidholdt Jr EM, Boone JM. The essential physics of medical imaging. 3rd ed. Lippincott Williams and Wilkins; 2012. 9. Kanal KM, Krupinski E, Berns EA, Geiser WR, Karellas A, Mainiero MB, et al. ACR-AAPM-SIIM practice guideline for determinants of image quality in digital mammography. J Digit Imaging 2013;26:10e25. 10. Pisano ED, Gatsonis C, Hendrick E, Yaffe M, Baum JK, Acharyya S, et al. Diagnostic performance of digital versus film mammography for breastcancer screening. N Engl J Med 2005;353(17):1773e83. 11. Skaane P, Young K, Skjennald A. Population-based mammography screening: comparison of screen-film and full-field digital mammography with soft-copy reading e Oslo I study. Radiology 2003;229:877e85. 12. Skanne P, Skjennald A. Screen-film mammography versus full-field digital mammography with soft-copy reading: randomized trial in a populationbased screening program e Oslo II study. Radiology 2004;232:197e204. 13. Vinnicombe S, Pinto Pereira SM, McCormack VA, Shiel S, Perry N, dos Santos Silva IM. Full-field digital versus screen-film mammography: comparison within the UK breast screening program and systematic review of published data. Radiology 2009;251(2):347e58. 14. Del Turco MR, Mantellini P, Ciatto S, Bonardi R, Martinelli F, Lazzari B, et al. Full-field digital versus screen-film mammography: comparative accuracy in concurrent screening cohorts. AJR Am J Roentgenol 2007;189:860e6. 15. Hambly NM, McNicholas MM, Phelan N, Hargaden GC, O'Doherty A, Flanagan FL. Comparison of digital mammography and screen-film mammography in breast cancer screening: a review in the Irish breast screening program. AJR Am J Roentgenol 2009;193:1010e8. 16. Seradour B, Heid P, Esteve J. Comparison of direct digital mammography, computed radiography and film-screen in the French National Breast Cancer Screening Program. AJR Am J Roentgenol 2014;202:229e36. 17. Yaffe MJ. Measurement of mammographic density. Breast Cancer Res 2008;10: 209. 18. Fischer U, Baum F, Obenauer S, Luftner-Nagel S, Heyden D, Vosshenrich R, et al. Comparative study in patients with microcalcifications: full-field digital mammography vs. screen-film mammography. Eur Radiol 2002;12:2679e83. 19. Neal CH, Coletti MC, Joe A, Jeffries DO, Helvie MA. Does digital mammography increase detection of high-risk breast lesions presenting as calcifications? AJR Am J Roentgenol 2013;201:1148e54. 20. Hofvind S, Skaane P, Elmore JG, Sebuødegård S, Hoff SR, Lee CI. Mammographic performance in a population-based screening program: before, during and after the transition from screen-film to full-field digital mammography. Radiology 2014;272(1):52e62. 21. Lee W. Cancer detection rates before and after conversion to digital. In: Presented at BreastScreen Australia Conference; 2014. Melbourne. 22. Evans J. Digital versus film: the impact of digital mammography on cancer detection rates in BreastScreen Victoria. In: Presented at BreastScreen Australia Conference; 2014. Melbourne. 23. Kirwan CC. Breast cancer screening: what does the future hold? Br Med J 2013;346. f87. 24. Heggie JCP, McLean ID, Herley J, Thomson FJ, Grewal RK, Diffey J, et al. ACPSEM position paper: recommendations for a digital mammography quality assurance program V3.0. Available from: http://www.ranzcr.edu.au/quality-a-safety/ radiology/practice-quality-activities/mqap; 2012 [Accessed 14/7/15]. 25. Keavey E, Phelan N, O'Connell AM, Flanagan F, O'Doherty A, Larke A, et al. Comparison of the clinical performance of three digital mammography systems in a breast cancer screening programme. Br J Radiol 2012;85(1016): 1123e7. 26. Chiarelli AM, Edwards SA, Prummel MV, Muradali D, Majpruz V, Done SJ, et al. Digital compared with screen-film mammography: performance measures in

None. Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

8

J.L. Diffey / Radiography xxx (2015) 1e9

27.

28. 29.

30. 31.

32.

33.

34.

35.

36. 37.

38.

39. 40. 41.

42. 43. 44.

45.

46. 47.

48.

49.

50.

51.

52.

53.

concurrent cohorts within an organized breast screening program. Radiology 2013;268(3):684e93. Schueller G, Riedl CC, Mallek R, Eibenberger K, Langenberger H, Kaindl E, et al. Image quality, lesion detection, and diagnostic efficacy in digital mammography: full-field digital mamography versus computed radiography-based mammography using digital storage phosphor plates. Eur J Radiol 2008;67: 487e96. Baldelli P, McCullagh J, Phelan N, Flanagan F. Comprehensive dose survey of breast screening in Ireland. Radiat Prot Dosim 2011;145:52e60. Yaffe MJ, Bloomquist AK, Hunter DM, Mawdsley GE, Chiarelli AM, Muradali D, et al. Comparative performance of modern digital mammography systems in a large breast screening program. Med Phys 2013;40(12):121915. http:// dx.doi.org/10.1118/1.4829516. Lazzari B, Belli G, Gori C, Del Turco MR. Physical characteristics of five clinical systems for digital mammography. Med Phys 2007;34(7):2730e43. Van Engen R, Young KC, Bosmans H, Thijssen M. Addendum on digital mammography Version 1, November 2003. Addendum to chapter 3 of the European guidelines for quality assurance in mammography screening. 3rd ed. European Reference Organisation for Quality Assured Breast Screening and Diagnostic Services; 2001. Young KC, Oduko JM. Evaluation of Kodak DirectView mammography computerised radiography system, NHSBSP report 0504. NHS Cancer Screening Programmes; 2005. Young KC, Oduko JM. Technical evaluation of the Kodak DirectView mammography computerised radiography system using EHR-M2 plates, NHSBSP report 0706. NHS Cancer Screening Programmes; 2007. Bouwman R, Young K, Lazzari B, Ravaglia V, Broeders M, van Engen R. An alternative method for noise analysis using pixel variance as part of quality control procedures on digital mammography systems. Phys Med Biol 2009;54: 6809e22. BreastScreen Australia. BreastScreen Australia national accreditation standards: BreastScreen Australia quality improvement program. Canberra: Australian Government Department of Health and Ageing; 2008. RANZCR. Guidelines for quality control testing for digital (CR DR) mammography V3. Royal Australian and New Zealand College of Radiologists; 2012. Niklason LT, Christian BT, Niklason LE, Kopans DB, Castleberry DE, OpsahlOng BH, et al. Digital tomosynthesis in breast imaging. Radiology 1997;205(2): 399e406. Diekmann F. Digital breast tomosynthesis and breast CT. In: Bick U, Diekmann F, editors. Digital mammography. Berlin: Springer-Verlag Heidelberg; 2010. p. 199e209. Dobbins JT. Tomosynthesis imaging: at a translational crossroads. Med Phys 2009;36(6):1956e67. Helvie MA. Digital mammography imaging: breast tomosynthesis and advanced applications. Radiol Clin N Am 2010;48:917e29. Protocol for the Quality Control of the physical and technical aspects of digital breast tomosynthesis systems. Draft Version 0.15. European Reference Organisation for Quality Assured Breast Screening and Diagnostic Services; January 2014. Smith A. Full field breast tomosynthesis. Hologic White Paper W-LM-TOMO; March 2003. Mertelmeier T, Speitel J, Frumento C. 3D breast tomosynthesis e intelligent technology for clear clinical benefits. Siemens White Paper. SenoClaire GE. Breast tomosynthesis clinical and non-clinical information supplement to operator manual. 5415896-3-1EN, Revision 1, 2013e2014. General Electric Medical. Fuji Amulet Innovality. Available from: http://www.fujifilm.eu/eu/products/ medical-systems/products/digital-mammography/p/amulet-innovality/ [accessed 29.03.15]. Smith A. Design considerations in optimising a breast tomosynthesis system. Hologic White Paper WP-00046. Ciatto S, Houssami N, Bernardi D, Caumo F, Pellegrini M, Brunelli S, et al. Integration of 3D digital mammography with tomosynthesis for population breast-cancer screening (STORM): a prospective comparison study. Lancet Oncol 2013;14:583e9. Rose SL, Tidwell AL, Ice MF, Nordmann AS, Sexton R, Song R. A reader study comparing prospective tomosynthesis interpretations with retrospective readings of the corresponding FFDM examinations. Acad Radiol 2014;21(9):1204e10. Skanne P, Bandos AI, Gullien R, Eben EB, Ekseth U, Haakenaasen U, et al. Comparison of digital mammography alone and digital mammography plus tomosynthesis in a population-based screening program. Radiology 2013;267(1):47e56. Friedewald SM, Rafferty EA, Rose SL, Durand MA, Plecha DM, Greenberg JS, et al. Breast cancer screening using tomosynthesis in combination with digital mammography. JAMA 2014;311(24):2499e507. McCarthy AM, Kontos D, Synnestvedt M, Tan KS, Heitjan DF, Schnall M, et al. Screening outcomes following implementation of digital breast tomosynthesis in a general-population screening program. J Natl Cancer Inst 2014;106(11). dju316. Gennaro G, Toledano A, di Maggio C, Baldan E, Bezzon E, La Grassa M, et al. Digital breast tomosynthesis versus digital mammography: a clinical performance study. Eur Radiol 2010;20:1545e53. Teertstra HJ, Loo CE, van den Bosch MA, van Tinteren H, Rutgers EJ, Muller SH, et al. Breast tomosynthesis in clinical practice: initial results. Eur Radiol 2010;20:16e24.

54. Wallis MG, Moa E, Zanca F, Leifland K, Danielsson M. Two-view and singleview tomosynthesis versus full-field digital mammography: high- resolution X-Ray imaging observer study. Radiology 2012;262(3):788e96. 55. Skaane P, Bandos AI, Eben EB, Jebsen IN, Krager M, Haakenaasen U, et al. Tomosynthesis screening with synthetically reconstructed projection images: comparison with digital breast tomosynthesis with full-field digital mammographic images. Radiology 2014;271(3): 655e63. 56. Cavagnetto F, Taccini G, Rosasco R, Bampi R, Calabrese M, Tagliafico A. ‘Invivo’ average glandular dose evaluation: one-to-one comparison between digital breast tomosynthesis and full-field digital mammography. Radiat Prot Dosim 2013;157(1):53e61. 57. Olgar T, Kahn T, Gosch D. Average glandular dose in digital mammography and breast tomosynthesis. Fortschr Rontgenstr 2012;184(10):911e8. 58. Feng SSJ, Sechopoulos I. Clinical digital breast tomosynthesis system: dosimetric characterization. Radiology 2012;263(1):35e42. 59. Tromans C, Highnam R, Morrish O, Black R, Tucker L, Gilbert F, et al. Which women based on clinical benefits and dose should be considered for breast screening with tomosynthesis?. In: Presented at AAPM conference; 2014. 60. Tromans C, Highnam R, Morrish O, Black R, Tucker L, Gilbert F, et al. Patient specific dose calculation using volumetric breast density for mammography and tomosynthesis. In: Fujita H, Hara T, Maramatsu C, editors. Proceedings of the 12th International Workshop on digital mammography. Berlin: SpringerVerlag; 2014. p. 158e65. LNCS 8539. 61. Rafferty RA, Niklason L, Jameson-Meehan L. Breast tomosynthesis: one view or two?. In: Presented at RSNA annual meeting, Chicago IL; 2006. Session SSG01e04 Breast Imaging (digital tomosynthesis). 62. Beck N, Butler R, Durand M, Andrejeva L, Hooley R, Horvath L, et al. One-view versus two-view tomosynthesis: a comparison of breast cancer visibility in the mediolateral oblique and craniocaudal views. In: American Roentgen Ray Society annual meeting; April 2013. Washington DC. 63. Baker JA, Lo JY. Breast tomosynthesis: state of the art and review of the literature. Acad Radiol 2011;18(10):1298e310. 64. Lang K, Andersson I, Rosso A, Tingbery A, Timberg P, Zackrisson S. Performance of one-view breast tomosynthesis as a stand-alone breast cancer screening modality: results from the Malmo Breast Tomosynthesis Screening Trial, a population-based study. Eur Radiol 2015. http://dx.doi.org/10.1007/ s00330-015-3803-3. 65. Zuley ML, Guo B, Catullo VJ, Chough DM, Kelly AE, Lu AH, et al. Comparison of two-dimensional synthesized mammograms versus original digital mammograms alone and in combination with tomosynthesis images. Radiology 2014;271:664e71. 66. D'Orsi CJ, Sickles EA, Mendelson EB, Morris EA, et al. ACR BI-RADS® Atlas, breast imaging reporting and data system. Reston, VA: American College of Radiology; 2013. http://www.acr.org/Quality-Safety/Resources/BIRADS/About-BIRADS/ How-to-Cite-BIRADS. 67. Smith A. The use of breast tomosynthesis in a clinical setting. Hologic White Paper; 2014. 68. Dromain C, Balleyguier C. Contrast-enhanced digital mammography. In: Bick U, Diekmann F, editors. Digital mammography. Berlin: Springer-Verlag Heidelberg; 2010. p. 186e98. 69. Dromain C, Balleyguier C, Adler G, Garbay JR, Delaloge S. Contrast-enhanced digital mammography. Eur J Radiol 2009;69:34e42. 70. GE SenoBright CESM. Available from: http://www3.gehealthcare.com/en/ products/categories/mammography/contrast_enhanced_spectral_ mammography [accessed 29.03.15]. 71. Smith A. The principles of contrast mammography. Hologic White Paper WP00084-001; 2014. 72. Lobbes MBI, Lalji U, Houwers J, Nijssen EC, Nelemans PJ, van Roozendaal L, et al. Contrast-enhanced spectral mammography in patients referred from the breast cancer screening programme. Eur Radiol 2014;24:1668e76. 73. Fallenberg EM, Dromain C, Diekmann F, Renz DM, Amer H, Ingold-Heppner B, et al. Contrast-enhanced spectral mammography: does mammography provide additional clinical benefits or can some radiation exposure be avoided? Breast Cancer Res Treat 2014;146:371e81. 74. Francescone MA, Jochelson MS, Dershaw DD, Sung JS, Hughes MC, Zheng J, et al. Low energy mammogram obtained in contrast-enhanced digital mammography (CEDM) is comparable to routine full-field digital mammography (FFDM). Eur J Radiol 2014;83:1350e5. 75. Jochelson MS, Dershaw DD, Sung JS, Heerdt AS, Thornton C, Moskowitz CS, et al. Bilateral contrast-enhanced dual energy digital mammography: feasibility and comparison with conventional digital mammography and MR imaging in women with known breast carcinoma. Radiology 2013;266:743e51. 76. Fallenberg EM, Dromain C, Diekmann F, Engelken F, Krohn M, Singh JM, et al. Contrast-enhanced spectral mammography versus MRI: initial results in the detection of breast cancer and assessment of tumour size. Eur Radiol 2014;24: 256e64. 77. Danielsson M. Spectral imaging using photon-counting technology. Sectra White Paper; 2010. 78. Molloi S. Spectral mammography using a photon counting detector provides a unique method for measurements of breast density. Philips Healthcare White Paper; 2014. 79. Fujifilm 3D mammography viewer. Available from: http://www.fujifilm.com/ products/medical/mammography/3d/ [accessed 29.03.15].

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007

J.L. Diffey / Radiography xxx (2015) 1e9 80. Getty DJ, D'Orsi CJ, Pickett RM. Stereoscopic digital mammography: improved accuracy of lesion detection in breast cancer screening. In: Krupinski EA, editor. Proceedings of the 9th International Workshop on digital mammography. Berlin: Springer-Verlag; 2008. p. 74e9. LNCS 5116. 81. D'Orsi CJ, Getty DJ, Pickett RM, Sechopoulos I, Newell MS, Gundry KR, et al. Stereoscopic digital mammography: improved specificity and reduced rate of recall in a prospective clinical trial. Radiology 2013;266(1):81e8. 82. Maidment AD, Bakic PR, Albert M. Effects of quantum nose and binocular summation on dose requirements in stereoradiography. Med Phys 2003;30(12):3061e71. 83. Koning breast CT e P130025. Available from: http://www.fda.gov/ medicaldevices/productsandmedicalprocedures/deviceapprovalsandclearances/ recently-approveddevices/ucm433495.htm [accessed 29.03.15]. 84. Koning breast CT. Available from: m1.koningcorporation.com/Default.aspx; [accessed 29.03.15]. 85. Kalender WA, Beister M, Boone JM, Kolditz D, Vollmar SV, Weigel MCC. Highresolution spiral CT of the breast at very low dose: concept and feasibility considerations. Eur Radiol 2012;22:1e8. 86. O'Connell AM, Karellas A, Vedantham S. The potential role of dedicated 3D breast CT as a diagnostic tool: review and early clinical examples. Breast J 2014;20(6):592e605. 87. Lindfors KK, Boone JM, Newell MS, D'Orsi CJ. Dedicated breast computed tomography: the optimal cross-sectional imaging solution? Radiol Clin N Am 2010;48:1043e54. 88. Wolfe JN. Breast patterns as an index of risk for developing breast cancer. AJR Am J Roentgenol 1976;126:1130e7. 89. Wolfe JN. Risk for breast cancer development determined by mammographic parenchymal pattern. Cancer 1976;37:2486e92. 90. Boyd NF, O'Sullivan B, Fishell E, Simor I, Cooke G. Mammographic signs as risk factors for breast cancer. Br J Cancer 1982;45:185e93. 91. Boyd NF, Dite GS, Stone J, Gunasekara A, English DR, McCredie MRE, et al. Heritability of mammographic density, a risk factor for breast cancer. New Engl J Med 2002;347:886e94. 92. Boyd NF, O'Sullivan B, Campbell JE, Fishell E, Simor I, Cooke G, et al. Bias and the association of mammographic parenchymal patterns with breast cancer. Br J Cancer 1982;45:179e84. 93. Vachon CM, van Gils CH, Sellers TA, Ghosh K, Pruthi S, Brandt KR, et al. Mammographic density, breast cancer risk and risk prediction. Breast Cancer Res 2007;9:217. 94. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, et al. Mammographic density and the risk and detection of breast cancer. New Engl J Med 2007;356: 227e36. 95. Byrne C. Studying mammographic density: implications for understanding breast cancer. J Natl Cancer Inst 1997;89:531e3. 96. News on Density Reporting Legislation. Available from: http://www. areyoudenseadvocacy.org/worxcms_published/news_page170.shtml [accessed 29.03.15]. 97. Gail MH, Brinton LA, Byar DP, Corle DK, Green SB, Schairer C, et al. Projecting individualised probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 1989;81:1879e86. 98. Claus EB, Risch N, Thompson WD. Autosomal dominant inheritance of earlyonset breast cancer: implications for risk prediction. Cancer 1994;73:643e51. 99. Ford D, Easton DF. The genetics of breast and ovarian cancer. Br J Cancer 1995;72:805e12. 100. Ford D, Easton DF, Stratton M, Narod S, Goldgar D, Devilee P, et al. Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in

101.

102.

103.

104.

105. 106. 107.

108.

109.

110. 111.

112.

113.

114.

115.

116.

117.

9

breast cancer families: the breast cancer linkage consortium. Am J Hum Genet 1998;62:676e89. Barlow WE, White E, Ballard-Barbash R, Vacek PM, Titus-Ernstoff L, Carney PA, et al. Prospective breast cancer risk prediction model for women undergoing screening mammography. J Natl Cancer Inst 2006;98(17):1204e12. Schousboe JT, Kerlikowske K, Loh A, Cummings SR. Personalizing mammography by breast density and other risk factors for breast cancer: analysis of health benefits and cost-effectiveness. Ann Int Med 2011;155:10e20. Evans DG, Warwick J, Astley SM, Stavrinos P, Sahin S, Ingham S, et al. Assessing individual breast cancer risk within the UK National Health Service Breast Screening Program: a new paradigm for cancer prevention. Cancer Prev Res 2012;5:943e51. Sergeant JC, Musa S, Wilson M, Evans DG, Howell A, Astley S. Breast density measurement for personalised screening. Breast Cancer Res 2012;14(Suppl. 1): 17. American College of Radiology. Breast imaging reporting and data System® (BIRADS®). Reston, VA: American College of Radiology; 1992. Byng JW, Boyd NF, Fishell E, Jong RA, Yaffe MJ. Automated analysis of mammographic densities. Phys Med Biol 1996;41:909e23. Smith JH. Prediction of the risk of breast cancer using computer vision techniques [PhD thesis]. Department of Medical Biophysics, University of Manchester UK; 1998. Diffey J, Hufton A, Astley S. A new step-wedge for the volumetric measurement of mammographic density. In: Astley SM, Brady M, Rose C, et al., editors. Proceedings of the 8th International Workshop on digital mammography. Berlin: Springer-Verlag; 2006. p. 1e9. LNCS 4046. Malkov S, Wang J, Kerlikowske K, Cummings SR, Shepherd JA. Single X-ray absorptiometry method for the quantitative mammographic measure of fibroglandular tissue volume. Med Phys 2009;36(12):5525e35. Highnam R, Brady M. Mammographic image analysis. Dordrecht: Kluwer Academic; 1999. Hartman K, Highnam R, Warren R, Jackson V. Volumetric assessment of breast tissue composition from FFDM images. In: Krupinski EA, editor. Proceedings of the 9th International Workshop on digital mammography. Berlin: SpringerVerlag; 2008. p. 33e9. LNCS 5116. Highnam R, Brady M, Yaffe MJ, Karssemeijer N, Harvey J. Robust breast composition measurement e VolparaTM. In: Marti J, Oliver A, Freixenet J, Marti R, editors. Proceedings of the 10th International Workshop on digital mammography. Berlin: Springer-Verlag; 2010. p. 342e9. LNCS 6136. Eng A, Gallant Z, Shepherd J, McCormack V, Li J, Dowsett M, et al. Digital mammographic density and breast cancer risk: a case-control study of six alternative density assessment methods. Breast Cancer Res 2014;16(5):439. Fenton JJ, Taplin SH, Carney PA, Abraham L, Sickles EA, D'Orsi C, et al. Influence of computer-aided detection on performance of screening mammography. New Engl J Med 2007;356:1399e409. Gilbert FJ, Astley SM, Gillan MGC, Agbaje OF, Wallis MG, James J, et al. Single Reading with computer-aided detection for screening mammography. N Engl J Med 2008;359:1675e1684Pubmed. Cole EB, Zhang Z, Marques HS, Hendrick RE, Yaffe MJ, Pisano ED. Impact of computer-aided detection systems on radiologist accuracy with digital mammography. AJR Am J Roentgenol 2014;203:909e16. CAD provides boost for digital breast tomosynthesis. Available from: http:// www.auntminnieeurope.com/index.aspx?sec¼ser&sub¼def&pag¼dis&Item ID¼609181 [accessed 29.03.15].

Please cite this article in press as: Diffey JL, A comparison of digital mammography detectors and emerging technology, Radiography (2015), http://dx.doi.org/10.1016/j.radi.2015.06.007