PRIME: Software-based scatter correction White Paper

White paper PRIME: Software-based scatter correction enabling gridless digital mammography Lowering dose without compromise in image quality siemens-healthineers.com/mammography SIEMENS Healthineers PRIME: Software-based scatter correction enabling gridless digital mammography PRIME: Software-based scatter correction enabling gridless digital mammography By Dejan Danilovic, PhD, and Astrid Hamann Contents What is dose in digital mammography and why is it important 2 Methods to evaluate dose and image quality in mammography 4 PRIME Technology: Software-based scatter correction enabling gridless digital mammography at lower dose 6 Technical and clinical validation of PRIME Technology 7 Conclusion 10 What is dose in digital mammography and why is it important In X-ray imaging, the term “dose” is employed in 1. The majority of women undergoing screening different contexts and has several definitions. From a are healthy and asymptomatic. patient’s standpoint, “dose” is typically taken to mean organ dose, which refers to the radiation energy 2. Low-energy (soft) X-rays are more readily absorbed absorbed and deposited per unit mass of human tissue. by the tissue than other forms of radiation energy. In mammography, Average Glandular Dose (AGD) is a 3. Glandular breast tissue has a relatively high decisive quantity describing radiation energy deposition. sensitivity to radiation-induced cancer, particularly The SI unit for dose is the gray (1 Gy = 1 J/1 kg). in young women. Figure 1 illustrates the main dose parameters in mammo- Reducing patient dose is always an aim when working graphy. An average breast is assumed to be composed ontechnical improvements in mammography, but has of 50% adipose tissue and 50% glandular tissue. Since to beweighed against the effects any such developments the risk of development of breast cancer in adipose tissue mayhave on image quality. Ideally, AGD should be as low is minimal, radiation dosimetry is concerned only with as reasonably achievable (ALARA) without compromising the dose deposited in the glandular tissue. AGD cannot image quality. be measured and is calculated approximately from the specific tube output and tube load, using scaling factors based on Monte Carlo simulations to account for the irradiated X-ray spectrum, the breast thickness, and the signal-to-noise transfer performance of the imaging Average Glandular Dose (AGD) is an approximation detector, often described in terms of detective quantum of organ (i.e. patient) dose. AGD should be as low efficiency (DQE). AGD is a frequently discussed issue in as reasonably achievable without compromising mammography, particularly in the context of screening, image quality. for the following reasons: 2 PRIME: Software-based scatter correction enabling gridless digital mammography X-ray tube Air kerma without backscatter Incident dose Compression panel Surface dose Air kerma with backscatter K Average glandular dose = organ dose Surface dose + conversion/correction factors Detector <- Detector dose Figure 1: Overview of the main dose parameters in mammography 1.1 X-ray scattering Mammography generally requires a high image This paper describes a novel approach to reduce quality for the detection of microcalcifications, masses, scattered radiation in the image, which permits the use distortions, and other subtle features indicating of a lower incident dose and results in a lower AGD than malignant tissue. X-ray scattering by the breast tissue in conventional anti-scatter solutions, without is a physical process degrading image quality [1]. compromising image quality. The scattered radiation can be modeled by a low- frequency and a high-frequency component. The low-frequency component reduces the relative image contrast by a smoothly varying additive offset to the intensity values in the image. The high-frequency X-ray scattering by the breast tissue is a physical component decreases the contrast-to-noise ratio (CNR) process degrading image quality. by the additive noise of stochastic nature. 3 PRIME: Software-based scatter correction enabling gridless digital mammography Methods to evaluate dose and image quality in mammography According to European guidelines, the total radiation per unit mass) is recorded for each thickness. AGD is dose incurred during a single mammogram should not calculated using the exposure factors selected by the exceed 2.5 mGy for a standard breast thickness of 4.5 cm automatic exposure control (AEC) for the given [2]. In full-field digital mammography (FFDM), the total PMMA thickness and the measured entrance surface AGD applied by a typical mammographic exposure is air kerma. The AGD values attained are then compared about 1 mGy per breast and lies below these regulations. to the acceptable and achievable maximum AGD levels With regard to quality assurance, AGD is used to control published by the European authorities. This method and compare the performance of different mammography was also used by Dance et al. [5] to compare breast systems. When measuring and comparing doses, both dose dose in tomosynthesis. and image quality measurements should be considered. Although AGD calculated in this way comes close to Three approaches are commonly used to compare dose the patient dose actually applied, the values are not and image quality in mammography: CDMAM phantom, completely consistent with each other. The applied PMMA phantom, and clinical data. patient dose is highly dependent on the AEC exposure parameters, which are adjusted according to the dense 2.1 CDMAM phantom parts of an imaged breast. Since the PMMA phantom is homogeneous, the effects of AEC do not impact the Contrast detail mammography (CDMAM) phantoms, applied dose and are not represented in the calculated illustrated in Figure 2, are commonly employed for image phantom dose values. Furthermore, PMMA is slightly quality assurance purposes in digital mammography and more dense than the compressed breast, necessitating are also used by the National Health Service (NHS) in the the use of conversion factors in calculations (e.g. 70 mm United Kingdom for image quality comparisons [3,4]. PMMA is equivalent to 90 mm compressed breast). A CDMAM phantom consists of gold disks of varying thicknesses and diameters arranged in a 16 x 16 matrix. To be able to assess image quality, a “threshold thickness” 0.50 -0.40 U.M.C. St. Radboud 0.20 is determined for the disks. This is the minimum diameter [mm] CDMAM-phantom Type : 3.4 0.31 Serial nr : 1071 0.16 0.25 thickness of the first gold disk of a certain diameter to 0. 13 0.20 0.10 0.16 0.13 0.10 be visible in an FFDM image. The dose needed to achieve gold thickness [um] 0.08 0.06 visibility for a particular disk in a scan can then be used 0.05 0.08 to compare the dose required for a given image quality 0.04 2.00 2.00 with different mammography systems. 0.06 0.03 1.60 However, this dose cannot be directly related to the applied dose during mammography exams, as it merely diameter [mm] represents a theoretical threshold value. The properties of a clinical image are dependent on many other factors, 10 0.31 0.20 0.25 0.36 0.50 0.71 1.00 1.42 gold thickness [um] 0.16 and the images themselves are intended to be suitable 90 0 for diagnostic use. As such, the results of a CDMAM Figure 2: CDMAM phantom of Artinis medical systems phantom measurement provide information about the used for image quality assessment in digital mammography. dose required for given image quality under technical conditions, but not about the dose applied to a patient in a clinical setting. 2.2 PMMA phantom To evaluate AGD in the course of quality assurance activities, European guidelines recommend polymethyl- methacrylate (PMMA) phantoms made up of stacked PMMA plates (Figure 3) [2]. They absorb radiation in a manner similar to real breast tissue. A PMMA phantom is imaged with different numbers of blocks in a stack, corresponding to thin and thick breasts, and the entrance air kerma (kinetic energy released Figure 3: Example of a PMMA phantom. 4 PRIME: Software-based scatter correction enabling gridless digital mammography 2.3 Clinical data The actual dose applied to a patient can be best approxi- Although this method is not directly standardized mated within the scope of clinical studies, optimally for image quality, it is possible to measure diagnostic with > 1000 cases [6–8]. AGD is recorded in the Digital accuracy using benchmarks such as the detection Imaging and Communications in Medicine (DICOM) rate and recall rate for mammography examinations. header of digital images, and it is therefore possible to In this way, dose can be standardized for a clinical analyze and evaluate large numbers of measurements. outcome rather than for a technical image quality. Equivalent populations should be selected for such analyses because patient population dose can differ across different screening regions. Three approaches used to compare dose are: The dose calculated during actual examinations CDMAM phantom, PMMA phantom, and clinical data includes the influence of breast density on the chosen (Table 1). AEC exposure parameters. CDMAM phantom PMMA phantom Clinical study Measured values Dose required to display AGD AGD the threshold thickness of gold disks Pro + Standardized for image + Breast thickness + Measures dose applied to quality included actual patients + AEC influence addressed + Standardized for diagnostic accuracy Con Standardized for image Homogenous material Not standardized for image – – – quality: Breast limits breast anatomy quality (no phantom) anatomy not taken effect on AEC – into account Large number of measurements (patients) required Assessment of actual – – applied dose + ++ Table 1: Summary of measurement techniques for comparing dose in digital mammography. 5 PRIME: Software-based scatter correction enabling gridless digital mammography PRIME Technology: Software-based scatter correction enabling gridless digital mammography at lower dose 3.1 Anti-scatter grid Conventionally, an anti-scatter grid is placed between a dedicated SBSC algorithm that enables gridless digital the breast and the detector to reduce the image mammography at lower dose, without compromising degrading effects of X-ray scattering. However, the grid image quality. PRIME Technology estimates a scatter field not only attenuates scattered radiation but also primary by using 2D scatter kernels [14,15] generated via Monte radiation reaching the detector: it lowers the detector Carlo simulations based on object properties, object dose (Figure 1) [9]. geometry, and acquisition parameters (tube voltage and The lower detector dose due to the attenuation of the filter/anode combination). Figure 4 shows a simplified grid has to be compensated for by a higher incident dose, illustration of the PRIME algorithm. which in turn increases AGD (Figure 1). Scatter correction takes only seconds longer than standard image post-processing in grid-based acquisition 3.2 Software-based scatter correction and will not affect the usual workflow during screening If mammography is performed without an anti-scatter mammography. grid, software-based scatter correction (SBSC) as a novel In 2013, PRIME Technology was 510k cleared by the U.S. strategy can reduce scattering effects [10–12]. SBSC Food and Drug Administration. It is optionally available algorithms estimate a low-frequency component of the on the Mammomat Inspiration FFDM systems (upgrade scattered radiation and subtract it from the measured possible) in all countries. Mammomat Inspiration FFDM image [13]. Effects of a high-frequency scatter systems (upgrade possible) in all countries. component are diminished indirectly by a larger amount of primary radiation reaching the detector without the grid. PRIME (Progressive Reconstruction Intelligently 3.3 PRIME algorithm Minimizing Exposure) algorithm corrects scatter and enables gridless digital mammography at lower dose, Siemens Healthineers developed PRIME (Progressive without compromising image quality. Reconstruction Intelligently Minimizing Exposure), + Subtract scatter field – Segment breast Computed scatter field Database with scattering kernels 20 00 Input image Scatter field image Output image Figure 4: Simplified illustration of the PRIME algorithm for scatter correction [13]. PRIME Technology estimates and subtracts the low-frequency component of the scattered radiation in a process of convolution, using a database with 2D scatter kernels. Effects of the high-frequency component of the scattered radiation are reduced indirectly through a larger amount of primary radiation reaching the detector with gridless image acquisition. 6 PRIME: Software-based scatter correction enabling gridless digital mammography Technical and clinical validation of PRIME Technology 4.1 Radiation dose reduction: Phantom and clinical study by Fieselmann et al. [13] The phantom study assessed potential surface dose Findings: The dose reduction factor depended on the reduction with gridless acquisition and PRIME Technology ECBT (Figure 5). Generally, with increasing breast versus standard acquisition with a grid, while providing thickness, the scatter fraction [18] and the additive noise the same CNR. The study also examined a difference in increase, requiring more primary radiation (allowing less contrast-detail visibility [2] for gridless acquisition at this dose reduction) to achieve the same CNR as with a grid. reduced dose level compared with grid-based acquisition. Figure 6 illustrates a contrast-detail curve comparison Methods: To study dose reduction, PMMA phantoms for grid-based versus gridless acquisition for a phantom (20–70 mm thick, in 10 mm steps) were placed on the thickness of 30 mm. The paired curves were similar to detector of Siemens Mammomat Inspiration and were each other for all phantom thicknesses, suggesting that covered partly by a 0.2 mm aluminum foil. Sets of similar image quality can be obtained with both images were acquired using AEC with a grid, or using acquisition techniques despite dose reduction with varying exposure times without a grid for the same beam gridless acquisition. quality (i.e. tube voltage and filter/anode combination) as for AEC. The CNR was determined for two regions of The clinical part of the study investigated dose-reduced interest (1 cm2 squares), one placed inside and the gridless acquisition under realistic clinical conditions [13]. other outside the aluminum foil area [13]. Each PMMA Methods: In 75 women recalled for further diagnostic thickness was converted to an equivalent compressed mammograms after screening, two exposures were made 20 mm phantom thickness, 21 mm equivalent CBT breast thickness (ECBT) [16]. Dose reduction factors during the same compression phase, first in the AEC 2.83 w/ grid 2.00 were computed for gridless image acquisition (Figure 5). w/o grid + SBSC (31% dose reduction) mode with a grid, then without a grid and with a dose 1.42 In a study of contrast-detail visibility, images acquired reduced by decreasing the tube current-time product 1.00 0.71 0.50 with 20–70 mm thick phantoms composed of PMMA according to the results of the phantom study, while 0.36 blocks and CDMAM were evaluated using the automatic keeping all other acquisition parameters constant. The 0.25 scoring software CDCOM [17]. The contrast-detail images acquired without a grid were processed with 0.20 0.16 0.13 PRIME Technology before their finalization using standard 0.10 disc thickness [ μ m] curves were constructed for each phantom thickness. 0.08 mammographic image processing algorithms [19]. 0.06 0.05 0.04 0.03 0.06 0.08 0.10 0.13 0.16 0.20 0.25 0.31 0.40 0.50 0.63 0.80 1.00 1.25 1.60 2.00 disc diameter [mm] 35 30 mm phantom thickness, 32 mm equivalent CBT Measured values 2.83 .. ..... w/ grid O Fit 2.00 O 30 w/o grid + SBSC (35% dose reduction) 1.42 1.00 0.71 25 O. 0.50 0.36 0.25 0.20 20 0.16 0.13 Disk thickness [ μ m] 0.10 0.08 0.06 0.05 15 0.04 -3 Dose reduction factor [%] 0.03 0.06 0.08 0.10 0.13 0.16 0.20 0.25 0.31 0.40 0.50 0.63 0.80 1.00 1.25 1.60 2.00 10 Disk diameter [mm] ............... 20 30 40 50 60 70 80 90 60 mm phantom thickness, 75 mm equivalent CBT 2.83 Equivalent compressed breast thickness [mm] w/ grid 2.00 w/o grid + SBSC (11% dose reduction) 1.42 1.00 Figure 5: Surface dose reduction for gridless vs. grid-based Figure 6: Pair of contrast-detail curves (log-log plots) illustrating 0.71 0.50 acquisition ranged from 11% to 35% as a function of the ECBT similar image quality for grid-based acquisition and dose-reduced 0.36 (phantom study)* [13]. gridless acquisition plus PRIME Technology (phantom study) [13]. 0.25 0.20 0.16 0.13 0.10 disc thickness [ μ m] 0.08 0.06 0.05 0.04 * PRIME is used for a maximum breast thickness of 7 cm under compression 0.03 0.06 0.08 0.10 0.13 0.16 0.20 0.25 0.31 0.40 0.50 0.63 0.80 1.00 1.25 1.60 2.00 disc diameter [mm] 7 PRIME: Software-based scatter correction enabling gridless digital mammography In a subsequent blinded reading of image pairs side-by- side, five experienced radiologists compared the two Dose reduction with preserved image quality images on a 7-point scale (-3, -2, ... , +3) for several was confirmed in a first phantom and clinical study categories reflecting image quality. Based on biostatis- with PRIME Technology. tical considerations, non-inferiority of the gridless technique for any category was assumed for a mean rating > -0.3 points on the 7-point scale, which was statistically tested by a one-sided t-test with the thres- Category Mean SD P-value hold set to -0.3. Overall image quality 0.07 0.31 < 10-14 Findings: The women (56 ± 5 years old) had compressed Visibility of breast thickness (CBT) of 57 ± 15 mm (28 to 87 mm), Tissue near the breast edge 0.00 0.16 < 10-24 with the breast density ratings: ACR 1 (4%), ACR 2 (53%), ACR 3 (33%), ACR 4 (10%). Mammographic findings Structures in the pectoral 0.09 0.29 < 10-10 were: masses (57%), microcalcifications (41%), and muscle architectural distortions (20%) [13]. Noise 0.01 0.22 < 10-17 While the absolute AGD reduction with gridless Diagnostic certainty of acquisition plus PRIME Technology was similar across CBT ranges (Figure 7), the relative AGD reduction ranged Mass 0.12 0.30 0.12 from 12% (CBT: 85–94 mm) to 32% (CBT: 25–34 mm). Microcalcification 0.08 0.42 0.08 The relative dose reduction factors were comparable to that in the phantom study [13], with minor differences Architectural distortion 0.18 0.28 0.18 attributable to the limited, discrete choice of exposure values available in the manual exposure mode, and to Table 2: Results from the reading study comparing image pairs the variation of breast composition in the population. (clinical study) [13]. A positive mean value denotes a preference In principle, a higher dose is necessary for thicker breasts for dose-reduced gridless acquisition plus PRIME Technology over which – in combination with a smaller relative dose grid-based acquisition. Very low P-values indicate highly significant reduction for thicker breasts – leads to a similar absolute non-inferiority of the gridless option. reduction. The reading study results are shown in Table 2. P-values 4.2 Image quality improvement: Phantom study indicate highly significant non-inferiority of dose-reduced by Binst et al. [20] gridless acquisition combined with PRIME Technology. Methods: Images of 20–70 mm thick phantoms composed of PMMA blocks and CDMAM, acquired with 3.0 w/grid Siemens Mammomat Inspiration, were evaluated for w/o grid + SBSC contrast-detail visibility using the automatic scoring 2.5 software CDCOM [17]. As CDCOM was previously proven only for grid-based acquisition, the study validated 2.0 CDCOM for gridless acquisition [20]. 1.5 Images of the PMMA blocks were obtained in the AEC mode. For each phantom thickness, exposure parameters 1.0 were recorded. The same exposure parameters were then AGD [mGy] set manually for the CDMAM configurations of corres- 0.5 ponding equivalent thicknesses, while varying the tube H current-time product (mAs) to ± 60% and ± 156% of the 0 values with AEC [20]. 25–34 35–44 45–54 55–64 65–74 75–84 85–94 Compressed breast thickness [mm] Altogether, 960 CDMAM images were obtained for four scatter conditions (with a grid, without a grid, PRIME Figure 7: The absolute AGD reduction with gridless acquisitionplus Technology without a grid with PMMA used as “breast PRIME Technology was 0.26 ± 0.06 mGy (mean ± SD) andsimilar thickness”, and PRIME Technology without a grid with across CBT ranges (clinical study) [13]. ECBT used as “breast thickness”), six phantom thicknesses, five dose levels, and eight images for each condition. 8 PRIME: Software-based scatter correction enabling gridless digital mammography To assess AGD reduction for gridless acquisition 4.3 Clinical evidence: Large-scale clinical study (with and without PRIME Technology) versus standard by Larsen et al. [8] acquisition with a grid, AGD for each exposure was calculated using the method of Dance [16]. Cancer detection rate, recall rate, and cancer detection specificity are three commonly used performance Findings: PRIME Technology did not influence CDCOM indicators (benchmarks) for screening mammography. readings compared with gridless acquisition without A large population study presented by Dr. Lisbet Larsen PRIME Technology, but image homogeneity [2] was at the European Society of Radiology (ECR) annual improved (Figure 8), correcting the cupping artefacts, conference in 2015 compared these performance i.e. intensity bias. These artefacts can lead to serious indi-cators for grid-based acquisition and dose-reduced misclassifications, especially when intensity-based gridless acquisition combined with PRIME Technology. segmentation algorithms are used to classify tissue on images [21]. Methods: The study was conducted in Southern Denmark during 12 months before and 5 months after the complete change from conventional grid-based digital Image homogeneity is improved with PRIME mammography to gridless acquisition with PRIME Technology. Technology. The PRIME algorithm was incorporated in one and the same Siemens Mammomat Inspiration system. For study purposes, population characteristics and the true breast cancer prevalence rates were considered to be equivalent for the two time intervals associated with the 300 two screening methods. Scatter correction Findings: The turning point was December 2013, before 200 which 50,071 women received grid-based screening and after which 22,117 women underwent dose-reduced gridless screening combined with PRIME Technology. w/o grid The study results are summarized in Table 3. Pixel value 100 w/o grid w/o grid + PRIME (PMMA) Screening performance indicators were equivalent for w/o grid + PRIME (ECBT) 2000 grid-based and gridless screening. In particular, the IIII 0 cancer detection rate was identical (0.55%). There was no 0 500 1000 1500 Distance y-direction [mm] statistical difference in the recall rate (2.59% grid-based vs. 2.44% gridless with PRIME Technology) or specificity Figure 8: Improved image hozmogeneity with PRIME Technology (97.96% vs. 98.11%). has the potential to overcome cupping artefacts, i.e. intensity bias (phantom study) [20]. The Y-direction is parallel to the chest wall. Performance indicators for screening mammography were equivalent for PRIME (22,117 women) and conventional grid-based screening (50,071 women) in terms of the cancer detection rate (5.5/1000), recall rate (2.5%), and specificity (98%). Screening Number of Screening Recall rate Cancer detection Specificity method women screened interval [NR/NW] rate [NC/NW] [1-(NR-NC)/NW] Grid-based without 50,071 12 months 2.59% 0.55% 97.96% PRIME Technology (2.45–2.73) (0.49–0.62) (97.84–98.09) Gridless with PRIME Technology 22,117 5 months 2.44% 0.55% 98.11% (2.23–2.64) (0.45–0.64) (97.93–98.29) Table 3: Performance indicators (benchmarks) for screening mammography in Southern Denmark.[8] Data in brackets are 95% confidence intervals. Two-sided equivalence testing (α = 0.05) performed on the cancer detection rate and specificity showed no statistically significant difference between the two screening methods. (NC: Number of women with cancer, NR: Number of women recalled, NW: Number of women screened) 9 PRIME: Software-based scatter correction enabling gridless digital mammography Conclusion Especially in screening mammography, the radiation Technical and clinical validation of PRIME Technology dose to the woman should be as low as possible for the showed up to 30% dose reduction in comparison to required image quality. grid-based acquisition with Mammomat Inspiration, An anti-scatter grid usually removes X-ray scattering that depending on breast thickness. In a large study of 72,188 would have degraded the image quality, from the final women, the three commonly used performance image. This however requires an increase in incident and indicators for screening mammography (cancer detection thus in glandular dose, as part of the primary radiation is rate, recall rate, and cancer detection specificity) were removed with the anti-scatter grid and thus doesn’t reach equivalent for gridless dose reduced screening with the the detector. PRIME algorithm and conventional grid-based screening. The Siemens Healthineers PRIME (Progressive Reconstruc- PRIME is a highly valuable development in digital tion Intelligently Minimizing Exposure) algorithm for mammography, enabling a significant dose reduction scatter correction enables gridless image acquisition with without compromising image quality or clinical uncompromised image quality. performance. Glossary AEC Automatic exposure control NW Number of women screened AGD Average glandular dose PMMA Polymethylmethacrylate ALARA As low as reasonably achievable PRIME Progressive Reconstruction Intelligently Minimizing Exposure CBT Compressed breast thickness PRIME (PMMA) Calculation using PMMA thickness as input parameter “breast thickness” CDCOM Automated readout PRIME (ECBT) Calculation using ECBT as input parameter “breast thickness” CDMAM Contrast detail mammography SBSC Software-based scatter correction CNR Contrast-to-noise ratio SD Standard deviation DICOM Digital imaging and communications SI The international system of units in medicine ECBT Equivalent compressed breast thickness vs. Versus FFDM Full-field digital mammography w/ With (in figures) Kerma Kinetic energy released per unit mass w/o Without (in figures) NC Number of women with cancer NR Number of women recalled 10 PRIME: Software-based scatter correction enabling gridless digital mammography References [1] Bushberg JT, Seibert JA, Leidholdt EM, and Boone [10] Baydush AH and Floyd CE. Improved image quality in JM. The Essential Physics of Medical Imaging, digital mammography with image processing. edition. Lippincott Williams & Wilkins, Medical Physics 2000; 27(7): 1503-1508. Philadelphia, USA, 2012. [11] Nykanen K and Siltanen S. X-ray scattering in [2] Perry N, Broeders M, de Wolf C, Tornberg S, Holland full-field digital mammography. Medical Physics R, von Karsa L, and Puthaar E (editors). European 2003; 30(7): 1864-1873. Guidelines for Quality Assurance in Breast Cancer Screening and Diagnosis, 4th Edition. Office for [12] Tromans CE, Cocker M, and Brady SM. Digital scatter official publications of the European communities, removal for mammography and tomosynthesis Luxembourg, 2006. image acquisition. In: Proceedings of the 11th International Workshop on Digital Mammography [3] Strudley CJ, Warren LM, and Young KC. Technical (IWDM); 2012: 7361, 260-267. evaluation of Siemens Mammomat Inspiration digital breast tomosynthesis system. NHSBSP [13] Fieselmann A, Fischer D, Hilal G, Dennerlein F, Equipment Report 1306 (October 2013). Mertelmeier T, and Uhlenbrock D. Full-field digital mammography with gridless acquisition and [4] Strudley CJ, Looney P, Young KC. Technical software-based scatter correction: investigation evaluation of Hologic Selenia Dimensions digital of dose saving and image quality. In: Proceedings breast tomosynthesis system. NHSBSP Equipment of SPIE Medical Imaging 2013: Physics of Medical Report 1307 (October 2013). Imaging 2013; 8668: 86685Y. [5] Dance DR, Young KC, and van Engen RE. Estimation doi:10.1117/12.2007490 of mean glandular dose for breast tomosynthesis: [14] Ruhrnschopf E-P and Klingenbeck K. A general factors for use with the UK, European and IAEA framework and review of scatter correction breast dosimetry protocols. Physics in Medicine methods in x-ray cone-beam computerized and Biology 2011; 56(2): 453-471. tomography. Part 1: scatter compensation [6] Uhlenbrock D and Mertelmeier T. Comparison of approaches. Medical Physics 2011; 38(7): 4296- anode/filter combinations in digital 4311. mammography with respect to average glandular [15] Ruhrnschopf E-P and Klingenbeck K. A general dose. Fortschritte auf dem Gebiet der Röntgen- framework and review of scatter correction strahlen und der bildgebenden Verfahren methods in x-ray cone-beam computerized (Fortschr Röntgenstr) 2009; 181: 249-254. tomography. Part 1: scatter estimation [7] Bosmans H, Michielsen K, Lemmens K, and Jacobs J. approaches. Medical Physics 2011; 38(9): 5186- Mean glandular dose survey of 2D mammograms 5199. acquired with Siemens MAMMOMAT [16] Dance DR, Young KC, and van Engen RE. Further Inspiration system. Department of Radiology, factors for the estimation of mean glandular dose University Hospital of Leuven. A white paper using the United Kingdom, European and IAEA issued by Siemens, 2011. Available at: http:// breast dosimetry protocols. Physics in Medicine www.healthcare.siemens.de/mammography/ and Biology 2009; 54(14): 4361-4372. whitepapers [17] Young KC, Alsager A, Oduko JM, Bosmans H, [8] Larsen LB, Fieselmann A, Pfaff H, and Mertelmeier T. Verbrugge B, Geertse T, and van Engen R. Performance of gridless digital mammography Evaluation of software for reading images of the acquisition technique for breast screening: CDMAM test object to assess digital mammog- analysis of 22117 examinations. Presentation at raphy systems. In: Proceedings SPIE Medical ECR 2015. Abstract B-1025 (SS 1082). Imaging 2008: Physics of Medical Imaging 2008; [9] Salvagnini E, Bosmans H, Struelens L, and Marshall 6913: 69131C. NW. Quantification of scattered radiation in projection mammography: four practical methods compared. Medical Physics 2012; 39(6): 3167- 3180. 11 [18] Mertelmeier T and Bernhardt P. Scatter in digital On account of certain regional limitations of sales rights mammography: antiscatter grid versus slot- and service availability, we cannot guarantee that all scanning. In: Proceedings of SPIE Medical Imaging products/services/features included in this brochure are 2005: Physics of Medical Imaging 2005; 5745: available through the Siemens Healthineers sales 299-306. organization worldwide. Availability and packaging may [19] Zanca F, Jacobs J, Ongeval CV, Claus F, Celis V, vary by country and are subject to change without prior Geniets C, Provost V, Pauwels H, Marchal G, and notice. Bosmans H. Evaluation of clinical image The information in this document contains general processing algorithms used in digital mammog- descriptions of the technical options available and may raphy. Medical Physics 2009; 36(3): 765-775. not always apply in individual cases. [20] Binst J, Sterckx B, Bemelmans F, Cockmartin L, Van Siemens Healthineers reserves the right to modify the Peteghem N, Marshall N, and Bosmans H. design and specifications contained herein without prior Evaluation of automated CDMAM readings for notice. Please contact your local Siemens Healthineers non-standard CDMAM imaging conditions: sales representative for the most current information. gridless acquisitions and scatter correction. Radiation Protection Dosimetry 2015; 165(1-4): In the interest of complying with legal requirements 350-353. concerning the environmental compatibility of our products (protection of natural resources and waste [21] Yang X, Wu S, Sechopoulos I, and Fei B. Cupping conservation), we may recycle certain components artifact correction and automated classification where legally permissible. For recycled components we for high-resolution dedicated breast CT images. use the same extensive quality assurance measures as Medical Physics 2012; 39(10): 6397-6406. for factory-new components. Any technical data contained in this document may vary within defined tolerances. Original images always lose a certain amount of detail when reproduced. Siemens Healthineers Headquarters Siemens Healthineers AG Siemensstr. 3 91301 Forchheim, Germany Phone: +49 9191 18-0 siemens-healthineers.com Published by Siemens Healthineers AG · 16601 0425 online · © Siemens Healthineers AG, 2025