| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Departments of Diagnostic Radiology (T.M.B., D.O., G.R., K.L., S.S., U.R.B.) and Biometrics and Medical Informatics (S.K.), Otto-von-Guericke University, Leipziger Str 44, 39120 Magdeburg, Germany. From the 1997 RSNA scientific assembly. Received December 30, 1997; revision requested March 28, 1998; final revision received January 28, 1999; accepted April 14. Address reprint requests to T.M.B. (e-mail: Thomas .Bernhardt@medizin.uni-magdeburg.de).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Patterns of simulated interstitial lung disease and catheters were superimposed over an anthropomorphic chest phantom. Hard-copy images were generated at DSR (200-, 400-, and 600-speed), storage phosphor radiography (200- and 400-speed), and asymmetric (400-speed) and symmetric (200- and 400-speed) screen-film imaging. Surface doses were measured, and receiver operating characteristic analyses were performed.
RESULTS: No statistically significant differences were found between the detector systems with the same speeds for each interstitial pattern. Significantly poorer results were found at 600-speed DSR than at 200-speed DSR. Detection of catheters and nodules over high-attenuation areas was significantly worse with the symmetric screen-film system than with the other detectors. The surface dose with the DSR system, without a grid, was about 50% less than that of the other detector systems, with grids, at the same speed.
CONCLUSION: No significant difference was found in the diagnostic performance at DSR, storage phosphor radiography, and film-based radiography for simulated interstitial lung disease at corresponding speeds; there was a reduction in the surface dose of about 50% with the 400-speed DSR system.
Index terms: Lung, radiography, 60.11, 60.1215, 60.917 • Phantoms • Radiography, comparative studies, 60.11, 60.1215 • Radiography, digital, 60.1215 • Radiography, selenium detector • Radiography, storage phosphor • Receiver operating characteristic (ROC) curve
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Findings from recent studies (2) have shown equal or better results with digital detector systems with a 0.2-mm pixel size than with conventional screen-film systems. Storage phosphor radiography has improved considerably because of the recently developed type V imaging plates (Fuji Photo Film, Tokyo, Japan) and because of a qualitative improvement in quantum efficiency (3).
Interstitial lung disease has several patterns, and when the findings are subtle, they can be difficult to detect because of their low contrast and high spatial frequency. Therefore, the ability of radiologists to accurately recognize patterns of simulated interstitial lung disease with a new detector system is essential.
The aim of this phantom study was to compare the two digital detectors with symmetric or asymmetric screen-film systems in the detection of patterns of simulated interstitial lung disease, nodules, and catheters. The diagnostic image quality and the reduction in image quality at DSR with a dose reduction were also evaluated.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
Detector Systems
Curix HT 1000 Ortho-Medium screens/HTL films (35 x 35 cm; Agfa-Gevaert, Antwerp, Belgium) and Curix HT 1000 Ortho-Regular screens/HTL films (35 x 35 cm; Agfa-Gevaert) were used as symmetric screen-film systems. InSight VHC (Eastman Kodak, Rochester, NY) was used as an asymmetric screen-film system (35 x 43 cm).
A type V imaging plate (1,744 x 2,144 matrix, 10 bits per pixel, 0.2-mm pixel size, 35 x 43-cm format) was used for storage phosphor radiography (Digiscan 2 H; Siemens Medical Systems, Erlangen, Germany). Postprocessing was performed with a Magic View Workstation 1102 (Siemens Medical Systems). All images were obtained with the same x-ray system (Polydoros 50S generator and Multix CH/CPH, Vertix 2E wall-mounted Bucky unit; Siemens Medical Systems) at storage phosphor radiography and at symmetric or asymmetric screen-film imaging. A stationary detector system (Thoravision; Philips Medical Systems, Hamburg, Germany) was used at DSR (2,166 x 2,448 matrix, 12 bits per pixel, 0.2-mm pixel size). The hard-copy images obtained at storage phosphor radiography were photographed on film with a laser camera (Agfa-Gevaert; LR 3,300 laser imager matrix, 0.2-mm pixel size), with a hard-copy size of 32 x 39 cm. DSR images were printed with the laser imager, with a hard-copy size of 32.5 x 40.5 cm.
Detection Parameters
Detection parameters were not uniform, as each detector system was adapted according to the optimal recommended parameters. The source-to-object distance was 1.5 m at symmetric and asymmetric screen-film imaging and at storage phosphor radiography; the distance was 2.0 m at DSR.
A posteroanterior chest radiograph was obtained with the following parameters: (a) for symmetric screen-film imaging, 125 kVp, moving grid (ratio, 12:40), detector dose according to speed (200 or 400), and automatic exposure; (b) for asymmetric screen-film imaging, 121 kVp, moving grid (ratio 12:40), detector dose according to speed (400), and automatic exposure; (c) for storage phosphor radiography, 125 kVp, moving grid (ratio 12:40), detector dose according to speed (200 or 400), and automatic exposure; and (d) for DSR, 150 kVp, an air gap of 150 mm for scatter reduction, no use of an optional stationary grid, detector dose according to speed (200, 400, or 600), and automatic exposure.
Postprocessing for the Digital Detector Systems
Parameters adapted for daily routine use were used in a postprocessing procedure, as follows: A large filter kernel with a length of 3 cm and an enhancement factor of 0.5 were used with a sigmoid gradation curve at storage phosphor radiography according to the recommendation of Prokop et al (4). This procedure provided for sufficient assessment of regions with low and high attenuation and provided for the identification of low-contrast and subtle structures with high spatial resolution on one image.
Automatic postprocessing procedures were used as reported in previous studies of DSR (5–7). Two filters for unsharp mask filtering of the preprocessed images were used; high spatial frequencies for the modulation transfer function were enhanced by filtering with a nonlinear enhancement factor of 0.2 mm. Then, we selected a second processing procedure with an unsharp filter mask and a large filter kernel of 3-cm length. We used an enhancement factor of 0.8, a noise compensation of 0.5, and a sigmoid gradation curve for the postprocessing.
Dose Measurement
The left and right main chambers of the detector system were placed at the same position over the upper parts of the lungs in all detector systems. An ionization chamber unit (DALi; PTW, Freiburg, Germany) was fixed at the back of the phantom for the detection of the surface dose with each detector-speed system.
Evaluation
Four fellows (T.M.B., D.O., K.L., S.S.) in radiology and one attending physician (U.R.B.) in diagnostic radiology participated in the study. A standard film viewer was used for viewing of the hard-copy images. A total of 600 images (75 acrylic plates x eight detector-speed systems) were prepared in random order at different sittings. The interval between each sitting was at least 3 weeks to avoid learning effects. Hard-copy images of the phantoms without any simulated structures and examples of the simulated lesions were presented to the observer prior to each sitting.
A five-score confidence scale for the identification of the simulated structures was used, as follows: 1 was "not present," 2 was "probably not present," 3 was "indeterminate," 4 was "probably present," and 5 was "present."
Data and Statistical Analysis
A total of 36,000 observations (75 acrylic plates x 12 areas x five observers x eight detector-speed systems) were evaluated at receiver operating characteristic (ROC) analysis. The CORROC2 program (Metz C, University of Chicago, Ill) was used to calculate the ROC curve areas along with their SDs. The mean ROC curve values were calculated for each pattern and detector-speed system. The comparison of the different techniques was carried out by means of multifactorial analyses of variance in three steps, all of which were performed by using the statistical software SAS, version 6.12, PROC GLM (SAS institute, Cary, NC) (SAS/STAT user's guide, version 6. 4th ed. Vol 1. Cary, NC: SAS Institute; 1994).
In the first step, a three-factorial analysis was performed for all images that depicted a structure. The mean value from the five-score confidence scale was considered in relation to the different structures and to the detector-speed system as fixed factors and the image as a random factor. The analysis included a multiple comparison by means of the Tukey test with respect to the factor "detector-speed system." This test was used to control the type I experimental error rate.
As there were notable interactions between the detector-speed system and the kind of structure, a second step, a series of two-factorial analyses was conducted to consider the different structures (including no structure) one by one, with the mean value from the five-score confidence scale determined on the basis of the fixed factor "detector-speed system" and the random factor "image." Again, the different detector-speed systems were compared by using the Tukey multiple procedure.
In the third step of the analysis, the different structures were again considered separately. The single scores of the five observers were analyzed in three-factorial designs, with the factors "observer" (fixed repeated), "detector-speed system" (fixed), and "image" (random); the analysis included a multiple comparison of the factor "detector-speed system." For the repeated factor "observer," the results of the multivariate Wilks 
test were used. All tests were performed with a .05 probability of error.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
Moreover, there was no statistically significant difference between the storage phosphor radiographic system and the symmetric or asymmetric screen-film system at a speed of 400. When the radiologists compared DSR images acquired with the 200-speed system with those acquired with the 600-speed system, there was a statistically significant difference in their ability to detect abnormalities (P < .05). The performances of the asymmetric and symmetric screen-film systems were equivalent.
Nodules
Nodules over the lung.—There was no statistically significant difference in the radiologists' detection of this pattern with the detector systems at each speed and with a reduced surface dose at DSR (P > .05) (Fig 2c, 2d; Table). There was no statistically significant difference between imaging at a speed of 200 and imaging at a speed of 400 with all detector systems (P > .05). The dose reduction with the 600-speed system resulted in inferior image quality as compared with that of the 200-speed DSR system and with all other systems (P < .05).
Nodules over the mediastinum and diaphragm.—At DSR with the dose reduction, the ability of the radiologists to detect nodules in regions with higher attenuation was worse compared with their ability to detect nodules over the lung (P < .05) (Fig 2c, 2d). The imaging quality of digital detector systems was significantly superior to that of symmetric screen-film systems (P < .05), even at lower speeds (storage phosphor radiography at 400 speed vs symmetric screen-film imaging at a speed of 200 [P < .05]) (Figs 2c, 2d, 5a, 5b). The asymmetric screen-film system, compared with the symmetric screen-film system, was significantly superior in depicting nodules over the region of the mediastinum (P < .05). There was no difference between DSR and storage phosphor radiography at the same speeds.
|
|
|
|
|
|
Therefore, there was no statistically significant difference between DSR, storage phosphor radiography (with the recently developed type V imaging plates), and screen-film imaging with the same speed in the detection of simulated interstitial lung disease; however, there was a reduction of about 50% in the surface dose at DSR.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
There are differences on the chest radiographs obtained with the two digital systems because of the different detector techniques. Storage phosphor radiography is well established as a system for chest radiography and for cassette-based bedside studies. The DSR system is available only as a stationary system. DSR has a smaller signal-to-noise ratio and a larger detective quantum efficiency than has storage phosphor radiography. However, recent advances in storage phosphorradiography—type V imaging plates—have increased its quantum efficiency (3) and have remarkably improved it. Therefore, we did not find a significant difference between this storage phosphor radiographic system and DSR systems with the same speed.
We did not use a stationary grid with the DSR system, as we did with all of the other detector systems. The air gap that we integrated at DSR is sufficient for scatter reduction, according to Neitzel (11). Chotas et al (12) recommend the use of a grid for scatter reduction. Slone et al (13) demonstrated in a clinical trial that an additional antiscatter grid improves the image quality but doubles the radiation dose to the patient.
Clinical trials are necessarily limited because of the exposure of patients to x rays. It is possible to compare more than two detector systems or systems with different speeds only in phantom studies, which is why we chose this method. We focused on low-contrast structures and on subtle lesions, and we evaluated the assessment of simulated structures over the mediastinum and lungs.
In our study, the InSight VHC asymmetric screen-film system (Eastman Kodak) produced images of high quality in regions with high attenuation and maintained the ability to depict lesions over the lung. For all simulated patterns, chest radiographs obtained with the DSR system without a grid were equivalent to or superior to images obtained with all of the other detector systems at speeds of 200 and 400. Because of their larger latitudes, both digital detector systems were superior to the symmetric screen-film systems in the depiction of lesions over the mediastinum as areas with high attenuation. Our results concerning the dose reduction support the conclusion that chest radiographs should be obtained only with automatic exposure at a speed of 400, as we did not find a statistically significant difference between the 200- and 400-speed systems with all simulated patterns of interstitial lung disease. Nevertheless, a further dose reduction at a speed of 600 was not possible because of the inability to depict lesions over all areas of the thorax.
With our protocol, the surface dose at DSR is about 50% less than that of the other detector systems at the same speed.
By using the type V imaging plates at storage phosphor radiography, we improved the assessment of catheters, as compared with the results of Galanski et al (14), who used the previously established type IIIN imaging plates (Fuji Photo Film). In this respect, we were also able to demonstrate the improved performance of the InSight VHC asymmetric screen-film system (Eastman Kodak). The diagnostic performance of this system in depicting lesions over the mediastinum was superior to that of the symmetric screen-film system, without any disadvantage in the detection of lesions over the peripheral parts of the lungs.
In our study, there was no significant difference in the assessment of all patterns of simulated interstitial lung disease at speeds of 200 and 400 with each detector system. These results are not consistent with those of Kimme-Smith et al (15) in the detection of simulated nodules that range in diameter from 3 to 20 mm; in their study, a statistically significant decrease in the detection of nodules on computed radiographs was obtained with an exposure that was 20% lower. They, it should be noted, used different simulation materials and imaging plates.
Findings from recent studies (except for a couple studies [10,16]) have shown a valence for detection at storage phosphor radiography (with previous types of imaging plates) that is equivalent to and, over the mediastinum, even superior to that of the symmetric screen-film systems for some patterns of simulated interstitial lung disease (8,17,18). Leppert et al (10) found an improved assessment in the mediastinum, although detection of peripheral lesions was inferior. We demonstrated adequate detection over the mediastinum and peripheral areas with the type V imaging plates.
Findings from recent studies of DSR have shown an improvement on digital chest radiographs. In a clinical trial, Freund et al (19) obtained an equivalent detection of lung diseases at DSR and at symmetric screen-film imaging (200 speed), with a reduced surface dose of selenium. A preference for DSR systems over symmetric screen-film systems has been documented previously (20,21). Findings from a phantom study of linear patterns by Kehler et al (2) have established the superiority of the DSR system over the symmetric screen-film system and the two storage phosphor radiographic systems. Both storage phosphor radiographic detectors investigated by Kehler et al (2) demonstrated a smaller quantum efficiency than that of the type V imaging plates used in our study. There was no statistically significant difference in the detection of nodules in their study.
Nevertheless, despite the improvement in the type V imaging plates for storage phosphor radiography, DSR is still superior, with the equivalent surface dose, as our findings demonstrate. In a phantom study, Schaefer-Prokop et al (5) did not find a difference in the detection of nodules on images obtained at DSR, storage phosphor radiography, and asymmetric or symmetric screen-film imaging. However, DSR was superior to all other detector systems in depicting linear and micronodular patterns.
If the surface dose is reduced, an asymmetric screen-film system is slightly inferior to a symmetric screen-film system in the depiction of the simulated lesions. In contrast to findings from our study, no substantial advantages were found with the digital systems versus symmetric screen-film systems and with the asymmetric (400-speed) screen-film system versus symmetric (250-speed) screen-film system in the depiction of nodules over the mediastinum. However, Schaefer-Prokop et al (5) used type IIIN imaging plates for storage phosphor radiography, and they used different speeds at film-based radiography.
van Heesewijk et al (6) detected simulated interstitial micronodular patterns that were equivalent at DSR and symmetric screen-film imaging. Dose reductions to 45% and 65% did not result in a marked deterioration in the detection of this pattern, with their study design (7). These results did not agree with ours. The reason might be that we used a different study design for the simulation of the lesions. A similar diagnostic performance was shown with DSR and symmetric screen-film systems in 104 cases validated at computed tomography in a clinical trial (22). No additional grid was used at DSR in all of these studies.
The very large areas under the curves for the ground-glass pattern and catheters suggest that the depiction of these lesions was too easy. However, the results for these patterns did not differ from those for the other simulated interstitial patterns in our study.
Nevertheless, after all of these phantom studies, a clinical study needs to be performed to ensure that the results measured in phantom studies also have a clinical effect.
In summary, there was no statistically significant difference between DSR, storage phosphor radiography (with the recently developed type V imaging plates), and screen-film imaging with the same speed in the detection of simulated interstitial lung disease; however, there was a reduction of about 50% in the surface dose at DSR.
Practical application: Findings from this experimental study show that different patterns of simulated interstitial lung disease and catheters can be imaged with digital detector systems with no significant difference, when compared with film-based radiography. These findings support moving forward with the use of digital systems at routine chest radiography.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, T.M.B.; study concepts, T.M.B., U.R.B.; study design, T.M.B., D.O., U.R.B., G.R.; definition of intellectual content, T.M.B., D.O., U.R.B.; literature research, T.M.B., D.O.; experimental studies, T.M.B., D.O., U.R.B., G.R.; data acquisition, T.M.B., D.O., U.R.B., K.L., S.S.; data and statistical analyses, T.M.B., S.K., U.R.B.; manuscript preparation, T.M.B., D.O., U.R.B.; manuscript editing, T.M.B., D.O., U.R.B., G.R.; manuscript review, T.M.B., D.O., U.R.B.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L J M Kroft, W J H Veldkamp, B J A Mertens, J-P A van Delft, and J Geleijns Dose reduction in digital chest radiography and perceived image quality Br. J. Radiol., December 1, 2007; 80(960): 984 - 988. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Korner, C. H. Weber, S. Wirth, K.-J. Pfeifer, M. F. Reiser, and M. Treitl Advances in Digital Radiography: Physical Principles and System Overview RadioGraphics, May 1, 2007; 27(3): 675 - 686. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. M. Kroft, W. J. H. Veldkamp, B. J. A. Mertens, J. P. A. van Delft, and J. Geleijns Detection of simulated nodules on clinical radiographs: dose reduction at digital posteroanterior chest radiography. Radiology, November 1, 2006; 241(2): 392 - 398. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Bacher, P. Smeets, L. Vereecken, A. De Hauwere, P. Duyck, R. De Man, K. Verstraete, and H. Thierens Image quality and radiation dose on digital chest imaging: comparison of amorphous silicon and amorphous selenium flat-panel systems. Am. J. Roentgenol., September 1, 2006; 187(3): 630 - 637. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Rapp-Bernhardt, T. M. Bernhardt, H. Lenzen, R. Esseling, F. W. Roehl, M. Schiborr, I. Theobald-Hormann, and W. Heindel Experimental Evaluation of a Portable Indirect Flat-Panel Detector for the Pediatric Chest: Comparison with Storage Phosphor Radiography at Different Exposures by Using a Chest Phantom Radiology, November 1, 2005; 237(2): 485 - 491. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. W. Venema, M. van Straten, G. J. den Heeten, T. M. Bernhardt, H. Lenzen, and W. Heindel Digital Radiography of the Chest: Reassessment of the High-Voltage Technique? * Dr Bernhardt and colleagues respond: Radiology, April 1, 2005; 235(1): 336 - 338. [Full Text] [PDF]
|
|
|
|
|
|
S. Metz, P. Damoser, R. Hollweck, R. Roggel, C. Engelke, K. Woertler, B. Renger, E. J. Rummeny, and T. M. Link Chest Radiography with a Digital Flat-Panel Detector: Experimental Receiver Operating Characteristic Analysis Radiology, March 1, 2005; 234(3): 776 - 784. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. Bernhardt, U. Rapp-Bernhardt, H. Lenzen, F. W. Roehl, S. Diederich, K. Papke, K. Ludwig, and W. Heindel Low-Voltage Digital Selenium Radiography: Detection of Simulated Interstitial Lung Disease, Nodules, and Catheters--A Phantom Study Radiology, September 1, 2004; 232(3): 693 - 700. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Rapp-Bernhardt, F. W. Roehl, R. C. Gibbs, H. Schmidl, U. W. Krause, and T. M. Bernhardt Flat-Panel X-ray Detector Based on Amorphous Silicon versus Asymmetric Screen-Film System: Phantom Study of Dose Reduction and Depiction of Simulated Findings Radiology, May 1, 2003; 227(2): 484 - 492. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Eisenhuber, A. Stadler, M. Prokop, M. Fuchsjager, M. Weber, and C. Schaefer-Prokop Detection of Monitoring Materials on Bedside Chest Radiographs with the Most Recent Generation of Storage Phosphor Plates: Dose Increase Does Not Improve Detection Performance Radiology, April 1, 2003; 227(1): 216 - 221. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Goo, J.-G. Im, H. J. Lee, M. J. Chung, J. B. Seo, H. Y. Kim, Y.-J. Lee, J.-W. Kang, and J. H. Kim Detection of Simulated Chest Lesions by Using Soft-Copy Reading: Comparison of an Amorphous Silicon Flat-Panel-Detector System and a Storage-Phosphor System Radiology, July 1, 2002; 224(1): 242 - 246. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||
10 mm) on the right and left sides. (b) Posteroanterior image obtained with a 200-speed symmetric screen-film system in the same region does not depict the right nodule and depicts the left nodule more poorly. Both a and b were digitized, and the region of interest was magnified for better visualization.
