Published online before print May 23, 2007, 10.1148/radiol.2441060977
(Radiology 2007;244:130-137.)
© RSNA, 2007
Detection of Simulated Microcalcifications in a Phantom with Digital Mammography: Effect of Pixel Size1
Sankararaman Suryanarayanan, PhD, MBA,
Andrew Karellas, PhD,
Srinivasan Vedantham, PhD,
Ioannis Sechopoulos, MS, and
Carl J. D'Orsi, MD
1 From the Winship Cancer Institute and Department of Radiology, Emory University School of Medicine, 1701 Uppergate Dr, Bldg C, Suite 5018, Atlanta, GA 30322 (S.S., A.K., S.V., I.S., C.J.D.); and Wallace H. Coulter Department of Biomedical Engineering at Georgia Institute of Technology and Emory University, Atlanta, Ga (S.S., A.K., I.S.). Received June 6, 2006; revision requested August 7; revision received September 8; accepted October 5; final version accepted November 15. Supported in part by National Institutes of Health (NIH) grants (RO1-CA88792 and RO1-EB002123) from the National Cancer Institute (NCI) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB). Supported in part by an infrastructure grant from the Georgia Cancer Coalition (GCC).
Address correspondence to A.K. (e-mail: akarell{at}emory.edu).
 |
ABSTRACT
|
|---|
Purpose: To evaluate the effect of pixel size on the detection of simulated microcalcifications in a phantom with digital mammography.
Materials and Methods: A high-spatial-resolution prototype imager that yields variable pixel size (39 and 78 µm) and a clinical full-field digital mammography (FFDM) system that yields a 100-µm pixel size were used. Radiographic images of a contrast-detail (CD) phantom were obtained to perform four-alternative forced-choice observer experiments. Polymethylmethacrylate was added to obtain phantom thicknesses of 45 and 58 mm, which are typical breast thicknesses encountered in mammography. Phantom images were acquired with both systems under nearly identical exposure conditions by using an antiscatter grid. Twelve images were acquired for each phantom thickness and pixel size (for a total of 72 images), and six observers participated in this study. Observer responses were used to compute the fraction of correctly detected disks. A signal detection model was used to fit the recorded data from which CD characteristics were obtained. Repeated-measures analyses with mixed-effects linear models were performed for each of the six observers. All statistical tests were two sided and unadjusted for multiple comparisons. A P value of .05 or less was considered to indicate a significant difference.
Results: Statistical analysis revealed significantly better CD characteristics with 39- and 78-µm pixel sizes compared with 100-µm pixel size for all disk diameters and phantom thicknesses (P < .001). Increase in phantom thickness degraded CD characteristics regardless of pixel size (P < .001).
Conclusion: On the basis of the conditions of this study, reducing pixel size below 100 µm with low imaging system noise enhances the visual perception of small objects that correspond to typical microcalcifications.
© RSNA, 2007
|
INTRODUCTION
|
|---|
The results of a large-scale clinical study involving a variety of digital mammography systems indicated that the overall diagnostic accuracy of full-field digital mammography (FFDM) as a means of screening for breast cancer was similar to that of screen-film mammography (1). However, the study findings also indicated that digital mammography was more accurate in women younger than 50 years, women with radiographically dense breasts, and pre- or perimenopausal women (1). Despite the advantages of FFDM over screen-film mammography, the relative performance of FFDM compared with that of screen-film mammography for the detection of specific types of features, particularly subtle microcalcifications, remains to be established. The large-area contrast properties of FFDM suggest that its ability to depict soft-tissue masses and architectural distortions probably exceeds that of screen-film mammography, but it is not quite as clear whether this holds true for the detection of microcalcifications. Study findings have demonstrated that the mammographic presentation of ductal carcinoma in situ is limited to microcalcifications in up to 72% of instances (2–5). Berg et al (6) found that 20% of biopsies for amorphous calcifications revealed malignant findings. These calcifications are at the threshold of visibility, and any parameter that improves their detection is important. The larger effective pixel size and the resulting lower spatial resolution of digital mammography compared with those of screen-film mammography warrant particular attention for the optimization of FFDM systems with respect to detectability of subtle microcalcifications.
Currently, there is no consensus on desirable pixel size or spatial resolution for digital mammography systems. Clinical amorphous silicon–based FFDM systems yield a pixel size of 100 µm (7,8), whereas clinical amorphous selenium–based FFDM systems yield a pixel size of 70 µm (9). Smaller pixel sizes for digital mammography are offered by a slot-scan system based on charge-coupled device technology (10) and a photon-counting detector system in a slot-scan geometry (11,12). In addition, smaller spatial sampling is offered by photostimulable storage phosphor technology (13).
Results of various studies point to the possible need for higher spatial resolution in digital mammography (14–17). Yamada and colleagues (16) concluded that high spatial resolution is required with digital mammography to successfully differentiate between microcalcifications. Ruschin et al (17) found improved performance in terms of microcalcification shape determination with pixel sizes of less than 100 µm. The pixel sizes yielded by current large-area, flat-panel digital mammography imagers range from 70 to 100 µm, but knowledge on the performance of smaller pixels is limited. However, current flat panel–based systems that use amorphous silicon or amorphous selenium detectors cannot be used to study the effect of small pixels because pixel sizes of approximately 50 µm or less are unavailable. The purpose of our study was to evaluate the effect of pixel size on the detection of simulated microcalcifications in a phantom with digital mammography.
|
MATERIALS AND METHODS
|
|---|
Clinical FFDM System
A clinical FFDM system (Senographe 2000D; GE Medical Systems, Milwaukee, Wis) was used to acquire images with a pixel size of 100 µm (7,8). This system consists of a columnar cesium iodide scintillator coupled to an amorphous silicon photodiode array, with a pixel size of 100 µm providing a 19 x 23-cm field of view.
Prototype FFDM System
A 16 x 24-cm prototype FFDM imager was developed based on charge-coupled device (CCD) technology (Fairchild Imaging, Milpitas, Calif) (18,19). The imager was constructed by assembling six solid-state monolithic 8 x 8-cm CCD modules, where each module was designed to be three-side buttable (ie, each CCD molecule can be joined to three other CCD molecules on three sides) and comprised 2048 x 2048 pixels. The imaging device could be operated in the 39- or 78-µm pixel modes with a unity fill factor. The CCD design incorporated a full-frame architecture. Each CCD module was directly coupled to a nontapering fiberoptic faceplate (Type 47A; Schott North America, Southbridge, Mass). A cesium iodide scintillator with approximate length and width of 16 and 24 cm, respectively, and a thickness of 150-µm deposited on an amorphous carbon substrate (Hamamatsu, Bridgewater, NJ) was placed on the fiberoptic faceplate. A foam-type material was placed over the scintillator substrate, and the complete assembly was pressure bonded with a detector cover plate. The detector and electronics components were enclosed in a cassettelike format and operated without liquid circulation for cooling.
Contrast Detail Phantom
A contrast-detail (CD) phantom (CDMAM 3.2; Fluke Biomedical, Cleveland, Ohio), originally developed by Thijssen and colleagues (20), was used as the test object in this study. This phantom has been widely used for image quality assessment in mammography. The phantom consisted of a thin aluminum base that contained circular gold disks that were logarithmically sized from 0.10 to 3.2 mm in diameter and from 0.05 to 1.6 µm in thickness. The disks were arranged in a matrix of squares such that within each square one disk was centrally placed and an additional disk was randomly placed at one of the four corners. Within each square, the central and corner disks had the same diameter and thickness. Along a row of squares, the disk thickness was constant and the diameter varied logarithmically; along a column, the diameter remained constant and the disk thickness varied logarithmically (21). In this study, additional polymethylmethacrylate was added such that the total phantom thicknesses were 45 and 58 mm. The main advantage of this phantom was the presence of a randomly spaced corner disk in each square that facilitated four-alternative forced-choice experiments. It has been shown recently that for mammographic features smaller than 1 mm (ie, microcalcifications), the quantum and electronic noise characteristics of the imaging system are overwhelmingly dominant compared with the anatomic structural noise of the breast (22). Thus, it is appropriate to perform CD analysis by using a phantom without the anatomic noise that is encountered in a mammographic image. This phantom has been used previously to compare CD characteristics of a screen-film mammography system with those of an FFDM system (23).
Image Acquisition
Images of the CD phantom were first acquired with the clinical FFDM system in the contrast-auto mode to emphasize image contrast. The compression paddle and antiscatter grid were used to simulate a clinical situation. The mammography system automatically selected kilovolt peak, tube current, and target-filter combinations for both phantom thickness conditions. Twelve images were acquired with 100-µm pixel resolution for each phantom thickness condition, and the raw images were selected for this study. The raw images were automatically bad-pixel and flat-field corrected by the clinical FFDM system. The prototype FFDM imager was used with an older mammography unit (Senographe DMR; GE Medical Systems) with a source-to-detector distance of 66 cm, which was identical to that of the clinical FFDM system. A reciprocating antiscatter grid with a grid ratio of 5:1 was used during image acquisition. The images were acquired with both mammography units and under nearly identical exposure conditions (Table 1). As with the clinical system, 12 images were acquired for each phantom thickness (45 and 58 mm) and pixel condition (39 and 78 µm). The images were dark subtracted and flat-field and bad-pixel corrected by an author (S.S.) who implemented a computer program. Disk diameters of 0.13–0.31 mm were used in this study. For each disk diameter, six disk thickness levels were selected such that disk perception ranged from "marginally" to "easily" perceivable.
View this table:
[in this window]
[in a new window]
|
Table 1. Mammography Technique Factors and Measured Imager Entrance Exposures at 66 cm with No Phantom in the Beam Path for Clinical and Prototype Mammography Units
|
|
Image Processing and Display
A graphical user interface was developed by an author (S.S.) with Interactive Data Language software (IDL 6.0; Research Systems, Boulder, Colo). The squares in the phantom images that corresponded to the disk diameters and thickness of interest were cropped before display by an author (S.S.) using a computer program. The dimensions of the region of interest (ROI) for the 100-, 78-, and 39-µm pixel sizes were 100 x 100, 130 x 130, and 260 x 260 pixels, respectively. All image ROIs were displayed in the center of a Digital Imaging and Communications in Medicine–calibrated clinical gray-scale flat-panel display system (Dome C5i; Planar Systems, Beaverton, Ore). No zooming was performed to match the ROI sizes. Since there were five disk diameters, six disk thicknesses, and 12 images acquired per condition, 360 image ROIs were displayed to the observer for each phantom thickness and pixel condition. Visual cues were generated by the display program at each of the four corners of the square where one of the corners contained the disk. Each displayed ROI was adjusted with an automatic contrast enhancement technique. To achieve this, histogram analysis was performed to compute the density function of each ROI, with the maximum value set to the maximum digital value in that ROI image. The digital value corresponding to the peak of the histogram (also measured in digital units) was obtained, and the look-up table of the image display was scaled such that the threshold was set to be greater than half this value (Fig 1). A reference image ROI that contained a high-contrast version of the disk in the center was displayed above the image ROI to provide the observer with information related to the detection task. The diameter of the disk in the reference image always matched that in the ROI image. However, none of the observers was aware of the disk diameters or the type of ROI being displayed during the observation sessions. An automated messaging feature that indicated to the observer when 25%, 50%, 75%, and 100% of the observation session was complete was implemented.
Observer Study
Six observers (four board-certified radiologists with specialization in mammography and two graduate students with experience in medical imaging) participated in this study. The radiologists had 4, 7, 14, and 34 years of experience in mammography. The graduate students had 3 and 4 years of experience in medical physics and radiology research. Two authors (C.J.D., I.S.) participated as observers in the study.
The study was divided into six sessions in random order. On a given day, each observer was expected to complete two sessions. It took three visits per observer to complete the study. In each session, observers independently reviewed 360 image ROIs; this took 25–30 minutes, on average. For the complete study across six sessions, each observer reviewed 2160 image ROIs (360 ROIs were reviewed in each of six sessions). Before the commencement of each session, observers were trained with a small random subset of ROI images (approximately 15–20 ROIs for each condition) that were extracted from the acquired images until they felt comfortable with the task. Since the type of image ROI, display sequence, and position of the disk in each ROI were all random, the chances of learning or memory effects were virtually eliminated. No restrictions were placed on viewing distance, but observers were not allowed to adjust the window or level settings of the ROI images. All observations were conducted in a dedicated darkened radiologic image perception room and randomized between observers to minimize any systematic effects. This was a forced-choice study; therefore, observers had to indicate the location of the corner disk with a mouse click and were encouraged to arrive at the best estimate in cases where the disk was not perceivable. Observer responses were automatically recorded as true-positive or false-positive events and used to compute the "percent correct detection."
Data Analysis
On the basis of recorded observer responses, percent correct detection (sensitivity) values were computed for each observer, disk diameter, and thickness. To analyze the percent correct detection data we used a signal detection that hypothesized a continuous decision variable internal to the observer with Gaussian probability density functions for the presence or absence of the disk (24,25). A maximum-likelihood algorithm was implemented, as described by Ohara et al (24), to fit the detection data. The CD characteristics were obtained at the 62.5% detection threshold, as this provided a threshold between chance (25%) and correct (100%) detection. For each observer and diameter, the perceived disk thickness corresponding to the 62.5% correct detection threshold point in the detection curve was computed. Linear interpolation of adjacent points was performed wherever applicable. CD characteristics were derived for all diameters, observers, and phantom thickness conditions. Finally, the corresponding CD data obtained by individual observers were averaged to yield average CD characteristics.
Statistical Analysis
Repeated-measures analyses with mixed-effects linear models were performed for the perceived disk thickness outcome, as measured by each of the six observers. A saturated model was fit and included the fixed main effects (phantom thickness, pixel size, and diameter); the two-way statistical interaction (phantom thickness by pixel size, phantom thickness by diameter, and pixel size by diameter); and the three-way interaction (phantom thickness by pixel size by diameter). Analyses were performed with a means model by using statistical software (SAS Proc Mixed, version 8; SAS Institute, Cary, NC), which provided separate estimates of perceived mean disk thickness by phantom thickness, pixel size, and disk diameter. A compound symmetry variance-covariance form in observer measurements was assumed for perceived disk thickness, and robust estimates of the standard errors of parameters were used to perform statistical tests and construct 95% confidence intervals. All statistical tests were two sided and unadjusted for multiple comparisons. A P value of .05 or less was considered to indicate a statistically significant difference.
|
RESULTS
|
|---|
Overall, the computed percent correct detection characteristics indicated a degradation in detection with increased pixel size (Fig 2) under the conditions investigated. The mean perceived disk thickness for the 45- and 58-mm-thick phantoms changed in significantly different ways across pixel size and disk diameter (P < .001 for the three-way interaction [phantom thickness by pixel size by disk diameter]; P = .30, P < .001, and P < .001, respectively, for the two-way interactions [phantom thickness by pixel size, pixel size by disk diameter, and phantom thickness by disk diameter]).
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a: Examples of percent correct detection characteristics obtained from one observer for disk diameters of (a) 0.31 and (b) 0.20 mm at various pixel sizes for a phantom thickness of 45 mm. Degradation in detection with increasing pixel size is observed. The solid lines are maximum likelihood estimated detection characteristics.
|
|
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b: Examples of percent correct detection characteristics obtained from one observer for disk diameters of (a) 0.31 and (b) 0.20 mm at various pixel sizes for a phantom thickness of 45 mm. Degradation in detection with increasing pixel size is observed. The solid lines are maximum likelihood estimated detection characteristics.
|
|
For the 45-mm-thick phantom, perceived disk thickness for both 39- and 78-µm pixel sizes was significantly different (better threshold contrast) compared with perceived disk thickness for the 100-µm pixel size at each disk diameter (Fig 3a, Table 2). Perceived disk thickness was worse for the 78-µm pixel size than for the 39-µm pixel size at disk diameters of 0.13 and 0.16 mm but not at disk diameters of 0.20, 0.25, and 0.31 mm (Fig 3a), which indicated superior performance of the 39-µm pixel at disk diameters of 0.13 and 0.16 mm (Table 2).
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a: CD characteristics obtained at 62.5% detection threshold after averaging data from six observers for (a) 45-mm and (b) 58-mm phantom thickness conditions at 39-, 78-, and 100-µm pixel sizes. Lower (better) threshold CD characteristics at 39- and 78-µm pixel sizes are observed. Error bars indicate 95% confidence intervals.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b: CD characteristics obtained at 62.5% detection threshold after averaging data from six observers for (a) 45-mm and (b) 58-mm phantom thickness conditions at 39-, 78-, and 100-µm pixel sizes. Lower (better) threshold CD characteristics at 39- and 78-µm pixel sizes are observed. Error bars indicate 95% confidence intervals.
|
|
For the 58-mm-thick phantom, perceived disk thickness for the 39- and 78-µm pixel sizes was significantly different compared with perceived disk thickness for the 100-µm pixel size (Fig 3b, Table 3). Perceived disk thickness was worse for the 78-µm pixel size than for the 39-µm pixel size at a disk diameter of 0.13 mm (Fig 3b). Perceived disk thickness for the 78-µm pixel size was significantly (P = .04) different compared with that for the 39-µm pixel size at a disk diameter of 0.16 mm. Mean perceived disk thickness was similar for the 78- and 39-µm pixel sizes at disk diameters of 0.20 and 0.25 mm (Table 3). The difference between the mean perceived disk thickness at a disk diameter of 0.16 mm was significant, but the magnitude of this difference was small. On average, perceived disk thickness for all pixel sizes was significantly better for the 45-mm-thick phantom than for the 58-mm-thick phantom (Fig 4) at each of the five disk diameters (Table 4).
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a: CD characteristics obtained at 62.5% detection threshold after averaging data from six observers for (a) 100-µm, (b) 78-µm, and (c) 39-µm pixel sizes at 45- and 58-mm phantom thickness conditions. Degradation in CD characteristics regardless of pixel size is observed. Error bars indicate 95% confidence intervals.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b: CD characteristics obtained at 62.5% detection threshold after averaging data from six observers for (a) 100-µm, (b) 78-µm, and (c) 39-µm pixel sizes at 45- and 58-mm phantom thickness conditions. Degradation in CD characteristics regardless of pixel size is observed. Error bars indicate 95% confidence intervals.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c: CD characteristics obtained at 62.5% detection threshold after averaging data from six observers for (a) 100-µm, (b) 78-µm, and (c) 39-µm pixel sizes at 45- and 58-mm phantom thickness conditions. Degradation in CD characteristics regardless of pixel size is observed. Error bars indicate 95% confidence intervals.
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Comparison of the Effect of Phantom Thickness (45 vs 58 mm) for Each Pixel Size at Various Disk Diameters
|
|
|
DISCUSSION
|
|---|
The use of CD methods in the evaluation of clinical imaging systems is widely accepted, and such psychophysical characterization of imaging systems provides information on the image quality and diagnostic value of a modality (26,27). The method used in our study provides a means to quantitatively assess perception data and characterize imaging performance. Various investigators have successfully used similar approaches to assess radiographic image quality (21,26–34). An important factor that motivates CD analysis is the inclusion of the human observer in the imaging chain; this is a critical consideration with all medical imaging systems (35). CD methods can be extremely useful in a precursor study before a clinical trial to determine system performance or compare the performance among different systems in a controlled manner. CD analysis may reveal strengths or deficiencies in a system that may need to be corrected before the onset of comprehensive clinical studies. The logistics and cost considerations present substantial barriers to the evaluation of specific parameters with clinical trials. Hence, studies such as this one can be instructive in gaining a quantitative understanding of performance before embarking on a redesign of mammography systems or beginning comprehensive clinical trials. Other automated methods (36,37) can be used in combination with human observers to analyze system performance.
We focused on quantitatively comparing imaging performance for detection of simulated microcalcifications between two systems at various pixel sizes. However, a combination of pixel size and system noise probably affected the CD characteristics. The four-alternative forced-choice method used here provided an effective means to obtain psychophysical measurements (38). In our investigation, the random signal location probably mitigated observer learning bias to a great extent compared with more simplistic CD experiments, in which the location of the signal was fixed. The superior CD performance of the prototype system at 39- and 78-µm pixel sizes compared with the clinical FFDM system at a 100-µm pixel size can be attributed to the better modulation transfer factor and detective quantum efficiency characteristics of the prototype imager, especially at higher spatial frequencies (18). However, there are other means to improve the conspicuity of small high-frequency structures, such as microcalcifications, with postacquisition image processing techniques. The CD performance of the prototype unit with 45- and 58-mm phantom thicknesses demonstrates the viability of such imaging architectures for digital mammography.
The significantly better CD performance with the 39-µm pixel size compared with the performance with the 78-µm pixel size for the 45-mm-thick phantom with disk diameters of 0.13 and 0.16 mm demonstrates the resolving capability of small pixel sizes under the image noise conditions investigated. Similar rationale applies to the better performance of the 39-µm pixel for the 58-mm-thick phantom at a disk diameter of 0.13 mm. The unexpected reversal of performance at a disk diameter of 0.16 mm for the 58-mm-thick phantom, wherein the 78-µm pixel performed better, was primarily due to one observer who exhibited a much higher (poorer) detection threshold for the 39-µm pixel. However, the magnitude of this difference was small. At larger disk diameters under similar system noise conditions, it appears that the benefit of small pixel size is reduced. The degradation in detection and CD characteristics between the 45- and 58-mm-thick phantoms could be attributed to the differences in spectra and the possible increase in radiographic scatter that could have affected the depiction of small low-contrast objects regardless of pixel size.
Our study had limitations. Although this was not a clinical study with human subjects, the results demonstrate the viability and provide insights into the perceptual performance characteristics of large-area high-spatial-resolution imagers for detection of microcalcifications in the breast. Furthermore, since the effect of breast anatomic noise on microcalcification detection is minor (22), the general trends observed in this study are clinically relevant. However, a carefully designed clinical study may be required before a decision is made on the final pixel size for a specific imaging technology for mammography. The charge-coupled device in this study was used as a convenient platform to demonstrate the potential of high-spatial-resolution and low-noise digital mammography in microcalcification detection. Ideally, use of the same imager with multiple pixel sizes would have been preferred; however, such an imager was not available. It should also be noted that the results of our study are applicable to two-dimensional projections, as presently used in mammography, and that studies need to be conducted to explore the spatial resolution and noise effects for applications such as tomosynthesis and dedicated computed tomography of the breast.
Practical application: On the basis of the findings of this study, reducing pixel size below 100 µm with low imaging system noise enhances the visual perception of small objects that correspond to typical microcalcifications. Alternatively, enhancement of the system modulation transfer function through improved scintillator and detector technologies can improve microcalcification visualization. We believe our results can help in assessing the performance of similar imaging modalities for radiographic imaging. The CD analysis provides a means to compare and identify key parameters that influence system performance with human observers. The method we described can be used before the onset of clinical trials to obtain insights into the performance capability of an imaging system for a specific task.
|
ADVANCE IN KNOWLEDGE
|
|---|
- The results of this study indicate superior performance of 39- and 78-m pixel sizes with a high-spatial-resolution prototype digital imager compared with the performance of 100-m pixel size with a clinical full-field digital mammography system.
|
IMPLICATION FOR PATIENT CARE
|
|---|
- On the basis of the conditions investigated in this study, it appears that high-spatial-resolution (pixel size less than 100 µm) and low-noise digital radiographic mammography systems could potentially improve the detection and visualization of microcalcifications, leading to earlier and more accurate diagnosis.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank Kirk A. Easley, MS, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, Ga, for assistance with statistical analysis.
|
FOOTNOTES
|
|---|
Abbreviations: CD = contrast detail • FFDM = full-field digital mammography • ROI = region of interest
The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the GCC, NCI, NIBIB, or NIH.
A.K. has a research collaboration with GE Corporate Research and Development through Emory University. In 2006, he participated in a GE advisory board meeting on breast cancer. The present study is not related to any GE-sponsored research.
Author contributions: Guarantor of integrity of entire study, S.S.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, all authors; experimental studies, all authors; statistical analysis, S.S.; and manuscript editing, all authors
|
References
|
|---|
- Pisano ED, Gatsonis C, Hendrick E, et al. Diagnostic performance of digital versus film mammography for breast-cancer screening. N Engl J Med 2005;353:1773–1783.[Abstract/Free Full Text]
- Ikeda DM, Andersson I. Ductal carcinoma in situ: atypical mammographic appearances. Radiology 1989;172:661–666.[Abstract/Free Full Text]
- Stomper PC, Connolly JL, Meyer JE, Harris JR. Clinically occult ductal carcinoma in situ detected with mammography: analysis of 100 cases with radiologic-pathologic correlation. Radiology 1989;172:235–241.[Abstract/Free Full Text]
- Stomper PC, Connolly JL. Mammographic features predicting an extensive intraductal component in early-stage infiltrating ductal carcinoma. AJR Am J Roentgenol 1992;158:269–272.[Abstract/Free Full Text]
- Kinkel K, Gilles R, Feger C, et al. Focal areas of increased opacity in ductal carcinoma in situ of the comedo type: mammographic-pathologic correlation. Radiology 1994;192:443–446.[Abstract/Free Full Text]
- Berg WA, Arnoldus CL, Teferra E, Bhargavan M. Biopsy of amorphous breast calcifications: pathologic outcome and yield at stereotactic biopsy. Radiology 2001;221:495–503.[Abstract/Free Full Text]
- Vedantham S, Karellas A, Suryanarayanan S, et al. Full breast digital mammography with an amorphous silicon-based flat panel detector: physical characteristics of a clinical prototype. Med Phys 2000;27:558–567.[CrossRef][Medline]
- Suryanarayanan S, Karellas A, Vedantham S. Physical characteristics of a full-field digital mammography system. Nucl Instrum Methods Phys Res A 2004;533:560–570.[CrossRef]
- Saunders RS Jr, Samei E, Jesneck JL, Lo JY. Physical characterization of a prototype selenium-based full field digital mammography detector. Med Phys 2005;32:588–599.[CrossRef][Medline]
- Tesic MM, Fisher Piccaro M, Munier B. Full field digital mammography scanner. Eur J Radiol 1999;31:2–17.[CrossRef][Medline]
- Aslund M, Cederstrom B, Lundqvist M, Danielsson M. Scatter rejection in scanned multislit digital mammography. In: Yaffe MJ, Flynn MJ, eds. Proceedings of SPIE: medical imaging 2004—physics of medical imaging. Vol 5368. Bellingham, Wash: International Society for Optical Engineering, 2004; 478–487.
- Hemdal B, Herrnsdorf L, Andersson I, Bengtsson G, Heddson B, Olsson M. Average glandular dose in routine mammography screening using a Sectra MicroDose mammography unit. Radiat Prot Dosimetry 2005;114:436–443.[Abstract/Free Full Text]
- Seibert JA, Boone JM, Cooper VN, Lindfors KK. Cassette-based digital mammography. Technol Cancer Res Treat 2004;3:413–427.[Medline]
- Cowen AR, Launders JH, Jadav M, Brettle DS. Visibility of microcalcifications in computed and screen-film mammography. Phys Med Biol 1997;42:1533–1548.[CrossRef][Medline]
- Rong XJ, Shaw CC, Johnston DA, et al. Microcalcification detectability for four mammographic detectors: flat-panel CCD, CR, and screen/film. Med Phys 2002;29:2052–2061.[CrossRef][Medline]
- Yamada T, Ishibashi T, Sato A, et al. Comparison of screen-film and full-field digital mammography: image contrast and lesion characterization. Radiat Med 2003;21:166–171.[Medline]
- Ruschin M, Hemdal B, Andersson I, et al. Threshold pixel size for shape determination of microcalcifications in digital mammography: a pilot study. Radiat Prot Dosimetry 2005;114:415–423.[Abstract/Free Full Text]
- Suryanarayanan S, Karellas A, Vedantham S, Onishi SK. High-resolution imager for digital mammography: physical characterization of a prototype sensor. Phys Med Biol 2005;50:3957–3969.[CrossRef][Medline]
- Suryanarayanan S, Karellas A, Vedantham S, Sechopoulos I. Theoretical analysis of high-resolution digital mammography. Phys Med Biol 2006;51:3041–3055.[CrossRef][Medline]
- Thijssen MAO, Biijerk KR, Lindeyer JM, Project: quality assurance in mammography (QAMAM). Nijmegen, the Netherlands: University Medical Centre Nijmegen.
- Veldkamp WJ, Thijssen MA, Karssemeijer N. The value of scatter removal by a grid in full field digital mammography. Med Phys 2003;30:1712–1718.[CrossRef][Medline]
- Burgess AE. Effect of detector element size on signal detectability in digital mammography. In: Flynn MJ, ed. Proceedings of SPIE: medical imaging 2005—physics of medical imaging. Vol 5745. Bellingham, Wash: International Society for Optical Engineering, 2005; 232–242.
- Suryanarayanan S, Karellas A, Vedantham S, Ved H, Baker SP, D'Orsi CJ. Flat-panel digital mammography system: contrast-detail comparison between screen-film radiographs and hard-copy images. Radiology 2002;225:801–807.[Abstract/Free Full Text]
- Ohara K, Doi K, Metz CE, Giger ML. Investigation of basic imaging properties in digital radiography. XIII. Effect of simple structured noise on the detectability of simulated stenotic lesions. Med Phys 1989;16:14–21.
- Aufrichtig R. Comparison of low contrast detectability between a digital amorphous silicon and a screen-film based imaging system for thoracic radiography. Med Phys 1999;26:1349–1358.[CrossRef][Medline]
- Veldkamp WJ, Kroft LJ, Mertens BJ, Geleijns J. Digital slot-scan charge-coupled device radiography versus AMBER and Bucky screen-film radiography: comparison of image quality in a phantom study. Radiology 2005;235:857–866.[Abstract/Free Full Text]
- Uffmann M, Schaefer-Prokop C, Neitzel U, Weber M, Herold CJ, Prokop M. Skeletal applications for flat-panel versus storage-phosphor radiography: effect of exposure on detection of low-contrast details. Radiology 2004;231:506–514.[Abstract/Free Full Text]
- Hall TJ, Insana MF, Harrison LA, Soller NM, Schlehr KJ. Ultrasound contrast-detail analysis: a comparison of low-contrast detectability among scanhead designs. Med Phys 1995;22:1117–1125.[CrossRef][Medline]
- Cohen G, Wagner LK, Amtey SR, Di Bianca FA. Contrast-detail-dose and dose efficiency analysis of a scanning digital and a screen-film-grid radiographic system. Med Phys 1981;8:358–367.[CrossRef][Medline]
- Aufrichtig R, Xue P. Dose efficiency and low-contrast detectability of an amorphous silicon x-ray detector for digital radiography. Phys Med Biol 2000;45:2653–2669.[CrossRef][Medline]
- Chotas HG, Ravin CE. Digital chest radiography with a solid-state flat-panel x-ray detector: contrast-detail evaluation with processed images printed on film hard copy. Radiology 2001;218:679–682.[Abstract/Free Full Text]
- Verdun FR, Denys A, Valley JF, Schnyder P, Meuli RA. Detection of low-contrast objects: experimental comparison of single- and multi–detector row CT with a phantom. Radiology 2002;223:426–431.[Abstract/Free Full Text]
- Cook LT, Insana MF, McFadden MA, Hall TJ, Cox GG. Comparison of the low-contrast detectability of a screen-film system and third generation computed radiography. Med Phys 1994;21:691–695.[CrossRef][Medline]
- Wang J, Anderson J, Lane T, Stetson C, Moore J. Contrast-detail characteristic evaluations of several display devices. J Digit Imaging 2000;13(2 suppl 1):162–167.[Medline]
- Wang J, Langer S. A brief review of human perception factors in digital displays for picture archiving and communications systems. J Digit Imaging 1997;10:158–168.[Medline]
- Chakraborty DP, Eckert MP. Quantitative versus subjective evaluation of mammography accreditation phantom images. Med Phys 1995;22:133–143.[CrossRef][Medline]
- Chan HP, Lo SC, Niklason LT, Ikeda DM, Lam KL. Image compression in digital mammography: effects on computerized detection of subtle microcalcifications. Med Phys 1996;23:1325–1336.[CrossRef][Medline]
- Burgess AE. Comparison of receiver operating characteristic and forced choice observer performance measurement methods. Med Phys 1995;22:643–655.[CrossRef][Medline]
TOP
ABSTRACT
INTRODUCTION
