BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
British Journal of Radiology (2003) 76, 478-482
© 2003 British Institute of Radiology
doi: 10.1259/bjr/67597156
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Obenauer, S
Right arrow Articles by Grabbe, E
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Obenauer, S
Right arrow Articles by Grabbe, E

Full Paper

Dose reduction in full-field digital mammography: an anthropomorphic breast phantom study

S Obenauer, MD K-P Hermann, PhD and E Grabbe, MD

Department of Radiology, Georg-August-Universität Göttingen, Robert-Koch-Str. 40, 37 075 Göttingen, Germany

Correspondence: Dr. med. Silvia Obenauer. Department of Radiology, Georg-August-Universität Göttingen, Robert-Koch-Str. 40, 37 075 Göttingen, Germany


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 References
 
The aim of this study was to evaluate the potential for radiation dose reduction by using other beam qualities in full-field digital mammography (FFDM) compared with screen–film mammography (SFM). FFDM was performed using an amorphous silicon detector with a caesium iodide scintillator layer (Senographe 2000D, GE, Milwaukee, USA). SFM was performed using a state-of-the-art conventional system (Senographe DMR, GE, Milwaukee, USA) with a dedicated screen–film combination. An anthropomorphic breast phantom with superimposed microcalcifications (50–200 µm) was used to evaluate the detectability of microcalcifications. Contact mammograms and magnification views (m=1.8) performed with both the digital and the screen–film system were compared. Images were exposed automatically. Molybdenum/Molybdenum (Mo/Mo) anode–filter combination, 28 kVp and 63 mAs were selected by the automatic optimization of parameters (AOP) of the conventional system. This exposure protocol (protocol A) was also used as baseline for the digital system. Dose reduction in digital mammography was achieved by using protocol B with Mo/Rh and 31 kVp and protocol C with Rh/Rh and 32 kVp. The detectability of microcalcifications was assessed by 3 experienced readers with a confidence level ranging from 1 to 5. A receiver operating characteristic (ROC) analysis was performed. In protocol A the area under the ROC-curve (Az) for contact views performed by the screen–film system was 0.64 and for those performed with the FFDM system 0.68. The Az values were 0.74 in protocol B and 0.65 in protocol C for the digital system. For the conventional and digital magnification views Az values were 0.71 and 0.79, respectively. For protocol B the Az value was 0.81 and for protocol C it was 0.76. There is no statistically significant difference in the Az values for the different protocols in digital mammography and no significant difference from the screen–film system. A potential for dose reduction by using other beam qualities seems to be possible with this digital system.


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 References
 
There has been a lot of debate about spatial resolution in digital mammography. Conventional screen–film mammography (SFM) systems reach a spatial resolution of approximately 15 lp mm-1. The maximum spatial resolution of the digital system considered here is about 5 lp mm-1 [1–4]. As structures in a mammogram have low contrast, contrast resolution is a very important imaging parameter that, in conjunction with spatial resolution, strongly affects detection of structures. Although both contrast resolution and spatial resolution can be measured objectively using phantoms, studies must still be performed to evaluate how the trade-off between these and other image quality parameters will affect the detectability of cancers [5–9]. The striking advantage of digital systems is the high dynamic range and the linear relationship between dose at the detector and signal intensity, as opposed to the sigmoid relationship between optical density and dose in screen–film systems. Digital image processing technology can display overexposed or slightly underexposed images with normal image quality. This capability has prompted a discussion about the possibility of dose reduction in digital mammography. As we have shown in previous studies, a dose reduction in digital mammography by lowering the mAs results in loss of microcalcification detectability due to an increase in noise [10].

This study compares the potential for patient dose reduction using other anode–filter combinations and higher tube voltages in full-field digital mammography (FFDM) compared with a state-of-the-art conventional screen–film system by analysing the effect on the detectability of microcalcifications in an anthropomorphic breast phantom.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 References
 
Mammography systems
The study was performed using an FFDM system (Senographe 2000D) and a conventional screen–film system (Senographe DMR), both General Electric Medical Systems, Milwaukee, USA. The digital system is built on the DMR platform. Both systems have a dual track X-ray tube with a molybdenum and a rhodium anode track and a 0.03 mm molybdenum and a 0.025 mm rhodium filter. The focal spot size is 0.3 mm for standard view mammograms, and 0.1 mm for magnification views. The digital detector is composed of a caesium iodide scintillator with an amorphous silicon detector. The pixel size is 100 µm. After exposure, the images are displayed on a high resolution monitor (2 k x 2.5 k) at the review workstation and evaluated as softcopies. The radiologists were told to adjust brightness and contrast freely and to zoom interactively at the monitor. For the conventional screen–film system, a dedicated mammographic screen–film combination (UM-MA film with UM-MA fine screen, Fuji Photo Film, Tokyo, Japan) was used.

Both mammography systems were checked in an acceptance test according to the German Standards.

Anthropomorphic breast phantom
An anthropomorphic breast phantom with superimposed microcalcifications was used to evaluate the detectability of microcalcifications. A breast specimen proven to be free of microcalcifications was embedded in a paraffin block and superimposed with an acrylic test plate containing 25 blind holes. Each countersunk hole was 2 mm deep and 8.5 mm in diameter. 13 of the 25 blind holes were chosen randomly and filled with a variable number of eggshell fragments of a defined size between 50 µm and 200 µm to simulate clusters of microcalcifications [4, 11]. The eggshell fragments were fixed in place with circular shaped adhesive tape. By rotating the plate, a set of four images with different patterns of microcalcifications was obtained, providing a total number of 100 test fields (Figure 1Go). Contact mammograms and magnification views (m=1.8) obtained with both the digital and the screen–film system were compared. Figure 2Go shows a digital X-ray of the breast phantom. The exposure parameters (anode, filter, kVp, mAs) of the automatic mode (AOP, automatic optimization of parameters) of the conventional system which were optimized for daily routine work were used as the base-line protocol (protocol A). The same parameters were used for protocol A with the digital system and result in an almost equal dose at the detector as at the cassette location of the conventional system. The average optical net density of conventional mammograms was 1.6.



View larger version (109K):
[in this window]
[in a new window]
  		Figure 1. Anthropomorphic breast phantom with acrylic test plate. The phantom thickness is 4.5 cm and the test plate size is 7 cm x 7 cm.

 


View larger version (182K):
[in this window]
[in a new window]
  		Figure 2. Digital image of the anthropomorphic breast phantom (28 kVp, Mo/Mo, 63 mAs).

 
The average glandular dose (AGD), i.e. the mean value of the absorbed dose in the glandular tissue, is commonly used for an estimation of the radiation risk. The AGD cannot be measured directly, but was calculated on the basis of the entrance surface air kerma (ESAK) free in air, i.e. without backscatter, and the conversion factor g, which converts ESAK to AGD, by the relationship: Go


The ESAK for each exposure was calculated from the recorded mAs and tube output (measured in mGy mAs-1 for both systems and all relevant anode–filter combinations and settings of the tube voltage with a 1 cm3 flat chamber type 77334 (PTW, Freiburg, Germany)), and corrected to the actual source-to-skin distance (SSD), according to the European Protocol on Dosimetry in Mammography [12]. The conversion factor g depends mainly on radiation quality, and on thickness and tissue composition of the breast. We assumed a 50% glandular–50% adipose composition of the breast and used the conversion factor g calculated for this composition and a Mo/Mo anode–filter combination. Dance et al [13] recently introduced a correction factor s to Equation (1)Go, which corrects for any difference from the standard g tabulation owing to the use of a different anode–filter combination. Although differences exist for different anode–filter combinations, their magnitude (up to about 6%) is small compared with the estimated accuracy (approximately ±25%) in determining the AGD. It is therefore justified to use only one set of conversion factors as a function of half value layers for all anode–filter combinations [12]. This dose calculation is not valid for magnification views because of a smaller SSD, a smaller field size and partial irradiation [14]. The decrease in SSD usually results in a small decrease in factor g because the relative depth dose decreases more rapidly as a function of depth at short SSDs. This is the effect of the inverse-square law. In general, the factor g decreases as the field size decreases because of the reduced scattered radiation for the smaller irradiated tissue volume. In contrast to contact mammography, the whole breast is not exposed in magnification views, this partial irradiation yields another g reduction. Using the tabulated factor g in magnification mammography will, therefore, overestimate the resulting AGD. The actual dose will depend on the actual volume of tissue exposed. However, in our study all the magnification views were carried out under the same geometrical conditions, that means for the same exposed breast volume. Therefore the relative patient dose values which are always expressed as a percentage of the corresponding dose to the screen–film system are valid (Table 1Go).


View this table:
[in this window]
[in a new window]
  		Table 1. Exposure parameters and the resulting doses as entrance surface air kerma (ESAK) and average glandular dose (AGD) for both mammographic systems (SFM, screen–film mammography; FFDM, full-field digital mammography) and all protocols. The relative dose values in percentages are always related to the associated dose of the screen–film system

 
A receiver operating characteristic (ROC) analysis was performed by 3 readers with a 5-point confidence rating scale (program ROCFIT, of Charles Metz IBM version 1.2). The detectability was expressed by the area under the ROC curve, Az. The statistical significance of the differences in the ROC curves for different systems or protocols was estimated by using the Student paired t test for the Az values.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 References
 
As the contrast is user-defined in digital images, it was optimized by the reader for the image as a whole or individually for separate areas of higher or lower parenchymal density. Exposure parameters and the resulting doses (ESAK and AGD) for both mammographic systems and all protocols are listed in Table 1Go. The ROC curves of the two systems for protocol A are shown in Figure 3Go indicating that the digital imaging technique represented the better diagnostic method with the curves running more closely to the left upper corner. The ROC curves of the digital system for the protocols A–C are plotted for contact and magnification views in comparison with the conventional system (Figure 4Go). In protocol A the area under the ROC-curve, Az, for contact views performed by the screen–film system was 0.64 and 0.68 for the FFDM system. The Az values were 0.74 in protocol B and 0.65 in protocol C for the contact views in digital mammography. For the conventional and digital magnification views Az values were 0.71 and 0.79, respectively. For protocol B the Az value was 0.81 and for protocol C it was 0.76. The Az values are plotted in Figure 5Go. The lower dose images did not result in worse detectability; the average Az was higher in digital than in conventional mammography, but did not achieve statistical significance.



View larger version (24K):
[in this window]
[in a new window]
  		Figure 3. The receiver operating characteristic curves for the conventional and the digital system for contact (m=1.1) and magnification (m=1.8) mammograms. SFM, screen–film mammography; FFDM, full-field digital mammography.

 


View larger version (20K):
[in this window]
[in a new window]
  		Figure 4. The receiver operating characteristic curves for the conventional system and the digital system for the 3 dose protocols. (a) Contact and (b) magnification views. SFM, screen–film mammography; FFDM, full-field digital mammography.

 


View larger version (16K):
[in this window]
[in a new window]
  		Figure 5. Dependence of Az on technology and dose protocol for contact mammograms and magnification mammograms. SFM, screen–film mammography; FFDM, full-field digital mammography.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
References
 
The major potential advantages of digital mammography systems are improved handling, processing, computer assisted diagnosis (CAD), communication and archiving [3, 6–9]. FFDM with a 100 µm pixel pitch flat-panel a-Si detector provides a nominal spatial resolution of 5 lp mm-1 as compared with a spatial resolution of more than 15 lp mm-1 for conventional SFM. The minimum pixel size needed for digital mammography is still being discussed [3, 5]. A lot of studies emphasize the importance of optimized image processing on detail visualization in digital mammography [1, 8, 15, 16]. Our experimental data suggest that the present FFDM system is at least equivalent, if not superior, to conventional SFM in the detection of microcalcifications. Therefore, the spatial resolution that is given by the pixel size of the imaging system may not be the appropriate measure to characterize the imaging capabilities of a digital mammographic system. The detection of an object is dependent on its contrast relative to the background and the noise level. Detecting small, low contrast objects requires a combination of good contrast performance and low noise. However, neither low noise nor high contrast alone allows easy identification of the object, the ratio of contrast to background noise characterizes object detectability. The best way to characterize the system performance for contrast, noise and spatial resolution is the detective quantum efficiency, DQE. This is the signal-to-noise ratio squared at the detector output over the signal-to-noise ratio squared of the input X-ray image. Another way to look at DQE is that it is proportional to the image quality divided by the dose. So it gives an idea of the image quality that is achievable at any given dose. Doubling the DQE results in the same signal-to-noise ratio ("image quality") at half the dose or 40% improvement in signal-to-noise ratio at the same dose.

The aim of our study was to evaluate the potential for dose reduction by using other beam qualities with the FFDM system. Using the same anthropomorphic phantom it was shown that lowering the mAs results in a diminished detectability of microcalcifications [10]. In that study the reduction of the Az values was proportional to the lowering of the tube loading, mAs, that means proportional to the dose at the detector. In the present study, therefore, dose at the detector was held constant and reduction of patient dose was achieved by varying the radiation quality. The effect on the detectability of microcalcifications was analysed. Our preliminary results suggest that such alterations, which did not change the dose at the detector, have advantages over lowering the tube current for dose reduction. These results must be verified in systematic clinical trials which are currently under way at our institution. The study of Hermann et al [17] demonstrated that the Rh anode with the Rh filter was the most common anode–filter combination selected by the automatic exposure control (AOP mode: STD) for 83.6% of the exposures surveyed with the digital mammography system Senographe 2000D. The Mo anode was selected for 16.4% of all exposures. For 3.9% it was used in combination with the Mo filter and for 12.5% with the Rh filter. The present study confirms the appropriateness of these anode–filter combinations. Dance et al [18] concluded that for SFM, Mo/Mo is the spectrum of choice for all but the thickest or most glandular breasts. In digital mammography, an alternative spectrum is preferable for breasts thicker than 2 cm.

A limitation of our study is that we evaluated the radiation exposure for only one standard thickness with small high contrast objects; different thicknesses and low contrast objects were not investigated. Hermann et al [16] showed in a phantom study that the relative dose reduction was equivalent for small and large thicknesses of the phantom when changing the anode–filter combination from Mo/Mo to Rh/Rh. However, the absolute reductions were higher for a thick (50 mm PMMA) than a thin (30 mm PMMA) phantom. The (low) contrast-detail detectability was evaluated in another study with the digital mammography system Senographe 2000D [10]. Of course, detectability is not the only quality to be examined. The characterization of microcalcifications is a very important issue and should not be compromised by dose reduction. Phantom and clinical studies must be performed to evaluate whether the characterization of microcalcifications is equally possible when radiation exposure to the patient is reduced.

Received for publication July 3, 2002. Revision received April 7, 2003. Accepted for publication April 23, 2003.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 

  1. Brettle DS, Ward SC, Parkin GJ, Cowen AR, Sunson HJ. A clinical comparison between conventional and digital mammography utilizing computed radiography. Br J Radiol 1994;67:464–8.[Abstract/Free Full Text]
  2. Cowen AR, Parkin GJ, Hawkridge P. Direct digital mammography image acquisition. Eur Radiol 1997;7:918–30.[CrossRef][Medline]
  3. Feig SA, Yaffe MJ. Digital mammography, computer-aided diagnosis, and telemammography. Radiol Clin North Am 1995;33:1205–30.[Medline]
  4. Funke M, Breiter N, Hermann KP, Oestmann JW, Grabbe E. Storage phosphor direct magnification mammography in comparison with conventional screen-film mammography—a phantom study. Br J Radiol 1998;71:528–34.[Abstract]
  5. Hermann KP, Hundertmark C, Funke M, von Brenndorff A, Grabbe E. Digital mammography in direct magnification technique using a large-area amorphous silicon X-ray detector. Fortschr Röntgenstr 1999;170:503–6.
  6. Jarlman O, Samuelsson L, Braw M. Digital luminescence mammography. Acta Radiologica 1991;32:110–3.[Medline]
  7. Newstead GM, Weinreb JC. Critical pathway for the future: MR imaging and digital mammography. Radiographics 1995;15:951–62.[Medline]
  8. Shtern F. Digital mammography and related technologies: a perspective from the national cancer institute. Radiology 1992;183:629–30.[Abstract/Free Full Text]
  9. Yaffe MJ, Rowlands JA. X-ray detectors for digital radiography. Phys Med Biol 1997;42:1–39.[CrossRef][Medline]
  10. Obenauer S, Hermann KP, Schorn C, Fischer U, Grabbe E. Full-field digital mammography: dose-dependent detectability of simulated breast lesions. Fortschr Röntgenstr 2000;172:1052–6.[CrossRef]
  11. Obenauer S, Hermann KP, Schorn C, Funke M, Fischer U, Grabbe E. Full-field digital mammography: a phantom study for the detectability of simulated microcalcifications. Fortschr Röntgenstr 2000;172:646–50.[CrossRef]
  12. European Protocol on Dosimetry in Mammography. Luxembourg: European Commission, 1996.
  13. Dance DR, Skinner CL, Young KC, Beckett JR, Kotre CJ. Additional factors for the estimation of mean glandular breast dose using the UK mammography dosimetry protocol. Phys Med Biol 2000;45:3225–40.[CrossRef][Medline]
  14. Liu B, Goodsitt M, Chan HP. Normalized average glandular dose in magnification mammography. Radiology 1995;197:27–32.[Abstract/Free Full Text]
  15. Pisano ED. Current status of full-field digital mammography. Radiology 2000;214:26–8.[Free Full Text]
  16. Hermann KP, Obenauer S, Grabbe E. Radiation exposure in full-field digital mammography with a flat-panel x-ray detector based on amorphous silicon in comparison with conventional screen-film mammography. Fortschr Röntgenstr 2000;172:940–5.[CrossRef]
  17. Hermann KP, Obenauer S, Marten K, Kehbel S, Fischer U, Grabbe E. Average glandular dose with amorphous silicon full-field digital mammography—Clinical results. Fortschr Röntgenstr 2002;174:696–9.[CrossRef]
  18. Dance DR, Thilander AK, Sandborg M, Skinner CL, Castellano IA, Carlsson GA. Influence of anode/filter material and tube potential on contrast, signal-to-noise ratio and average absorbed dose in mammography: a Monte Carlo study. Br J Radiol 2000;73:1056–67.[Abstract]



This article has been cited by other articles:


Home page
 RadiologyHome page
E. Samei, R. S. Saunders Jr, J. A. Baker, and D. M. Delong
Digital Mammography: Effects of Reduced Radiation Dose on Diagnostic Performance
Radiology, May 1, 2007; 243(2): 396 - 404.
[Abstract] [Full Text] [PDF]

Home page
 RadioGraphicsHome page
M. Mahesh
AAPM/RSNA Physics Tutorial for Residents: Digital Mammography: An Overview
RadioGraphics, November 1, 2004; 24(6): 1747 - 1760.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Obenauer, S
Right arrow Articles by Grabbe, E
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Obenauer, S
Right arrow Articles by Grabbe, E

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
BJR DMFR IMAGING  ALL BIR JOURNALS