| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Full paper |
1Department of Radiation Physics, Göteborg University, Sahlgrenska University Hospital, SE-413 45 Göteborg, Sweden, 2Department of Radiology, Sahlgrenska University Hospital, SE-413 45 Göteborg, Sweden, 3Department of Radiation Physics, Lund University, Malmö University Hospital, SE-205 02 Malmö, Sweden, 4Department of Diagnostic Radiology, Malmö University Hospital, SE-205 02 Malmö, Sweden and 5GSF-National Research Center for Enviroment and Health, D-857 64 Neuherberg, Germany,
Correspondence: Birgitta Lanhede, Department of Radiation Physics, Norrlands Universitetssjukhus, SE-901 89 Umeå, Sweden
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
A surprisingly low correlation was seen between ESD and important parameters such as tube voltage and speed of the film–screen system. Some explanations for this result might be that nominal, not audited, values for tube voltage and film–screen system speed were recorded and used in the statistical analysis, that optical density of the images was not quantified, or that film processing conditions were not controlled. It could also be explained by the confounding influence of other parameters and that the correlation between parameters themselves could not be considered in the analysis of the collected data.
This might give rise to the suggestion that the predicted impact of quality control measures cannot be demonstrated by such trials with real patients and under routine conditions. However, without the ability to predict in some way the outcome of such measures on trial results in terms of patient dose and image quality derived from the selection of particular radiographic factors, it is impossible to develop effective optimization strategies for diagnostic radiology.
The framework of the Commission of the European Communities (CEC) research project "Predictivity and optimisation in medical radiation protection" addressed these fundamental operational limitations in existing radiation protection mechanisms in diagnostic radiology. An attempt was made to establish relationships between the quality of the radiological information content of the image, the image-producing procedure and the dose to the patient. The project also examined the weakness and strength of the CEC quality criteria [2]. A trial on conventional X-ray film–screen examinations of chest and lumbar spine was performed at the University Hospitals in Göteborg and Malmö in 1996/97. Details of the trial performance, the determination of radiological image content in terms of image criteria scores and visual grading analysis scores, and their relationship with the image-producing procedure are described by Almén et al [4]. The part of the trial dealing with chest imaging is described here.
The objectives of the present study were:
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Radiographic equipment
All images were produced with a Siemens Polydoros 50s generator and a Siemens (Bi150/30/52R–100 s/n 4056 1/0,6) X-ray tube using the same X-ray stand, Siemens Vertix-E, with a Siemens Jk-1 465*465 mm ion-chamber (Siemens AG, Erlangen, Germany). All images were exposed with automatic exposure control using Kodak films (35 cm x 43 cm T MAT L), Kodak screens (Lanex 160/320) and Kodak cassettes (X-Omatic LW) (Kodak, Rochester, NY). The films were developed in a Kodak M8 film processor with an Agfa G 138 developer.
Exposure parameters
Four physical and technical parameters were studied, with two different values each: tube voltage (102 kVp and 141 kVp), nominal speed class of the screen (160 and 320), maximum density in the parenchyma (1.3 and 1.8) and the method for scatter reduction (air gap 30 cm with focus-to-film distance (FFD) of 390 cm, and a grid with ratio R=12, 40 lamellae cm-1 and FFD 150 cm). The total filtration was 3.7 mm Al for 102 kVp and 5.7 mm Al for 141 kVp.
Dose measurements
The air kerma (kerma in free air) at the entrance surface (Kair,e) was measured with a semiconductor detector (Solidose 300; RTI Electronics, Mölndal, Sweden) positioned directly on the subject in the lower part of the radiation field. The semiconductor detector was not sensitive to backscattered radiation. The detector was compared with an ionization chamber, traceable to BIPM (Bureau International des Poids et Mesures, France). Difference in response was found to be less than the uncertainty of the measurements.
Kair,e is not a quantity related to radiation risk. Therefore, a risk-related dose HGolem (a weighted equivalent dose similar to effective dose and based on a voxel phantom, Golem, segmented from whole body CT data of an adult male) was calculated from the air kerma values as described by Zankl et al [5]. The conversion between air kerma and this risk-related dose is: 
|
|
where fc(HGolem) is a conversion coefficient (Sv Gy-1) [5]. The conversion coefficients used are presented in Table 1.
|
Automatic exposure control was used. Measurements of optical density (the maximum density in a circular region of 2 mm diameter in the upper left corner of the left lung) were performed by one person using the samedensitometer for all images. Densities from 1.65 to 1.95 were accepted for the darker films and from 1.15 to 1.45 for the brighter films. All the identification tags on the films were masked and the radiographs were combined two by two for each volunteer. The pairs of radiographs were randomly mixed and batched ten by ten.
Quality criteria and quality scores
The European guidelines on quality criteria for chest images as presented in the 1990 version [2] were used in the trial, since the 1996 version had not been published at that time. These guidelines consist of image criteria regarding the specific image quality of reproduced structures as well as image criteria regarding positioning, degree of inspiration, etc. A total of 10 image quality criteria were used in the present study. Four of these (numbers 1 to 4 below) were identical to those in the 1990 version and deal with image quality, and six of them (numbers 5 to 10 below) were modified from the original four criteria (in the following referred to as modified quality criteria). This revision was proposed after discussions within a European group of expert radiologists before the trial. The following questions were at issue.
Does the X-ray film comply with the following image quality criteria?
Details to be sharply visualized in the parenchyma
Details to be reproduced in the mediastinum
Details to be sharply visualized in the costopleural junction
Evaluation
The evaluation procedure took place in a room with dimmed light. Each radiologist had his/her own light box, which was masked to the size of the observed images. The luminance of each light box was measured at five well defined points. The mean value of the five measured points varied between 1420 cd cm-2 and 3520 cd cm-2 for the seven light boxes used. For a specific light box, the maximum relative standard deviation of the five measured points was 9%.
All the chest radiographs were examined in two different ways. (1) The observers viewed the radiographs separately and stated whether the CEC image quality criteria were fulfilled. Both the original CEC criteria (numbers 1–4 above) and the modified criteria (numbers 5–10 above) were used. (2) The observers viewed the radiographs in pairs (radiographs from the same volunteer, of which one was a reference image) and compared the image quality criteria using visual grading analysis (VGA). Only the modified criteria (numbers 5–10 above) were used. In VGA, the observer stated whether the visibility of a given structure or region in the tested image was clearly inferior to (-2), slightly inferior to (-1), equal to (0), slightly better than (1) or clearly better than (2) the visibility of the same structure or region in the reference image.
Furthermore, to estimate intraobserver variance, all observers re-read the first batch at the end of the trial session.
An image criteria score (ICS) and a score derived from the VGA, the visual grading analysis score (VGAS), were calculated for each radiographic technique.
The ICS related to a particular technique was defined as:
|
|
where the summation is performed for the number of observers (o), images (i) and criteria (c), and No is the total number of observers, Ni is the total number of images for each radiographic technique and Nc is the total number of criteria for each image, i.e. the numerator in Equation (1) equals the number of criteria fulfilled and the denominator equals the total sum of image criteria for all observers and images.
ICS is normally calculated for all criteria used in the analysis. An ICS can also be calculated for the individual observers (ICSo). In this case No=1. Furthermore, it can be calculated for the individual quality criteria (ICSc), in which case Nc=1.
The VGAS related to a particular technique was defined as:
|
|
where RRo,i,c is the relative rating for a particular observer (o), image (i) and criterion (c). No, Ni and Nc are as described above. Since the actual relative rating is dependent on the quality of the reference image, a randomization of reference images as used here may cause biased results. To avoid possible asymmetry in the number of times a given system is chosen as reference, both images in the evaluated pair are actually given a rating; the observer rates the image compared with the reference image, then in the subsequent calculation process the reference image itself is given the same score with the opposite sign. Thus, RRo,i,c is calculated with a positive sign if the image was compared with the reference image, with a negative sign for the reference image itself. It is assumed that the grading for each pair would be of the same magnitude but with opposite sign if the compared and reference image changed places. With this kind of symmetry the relative score for a system will be compared with the average score of all systems.
The significance of differences between different systems was calculated using analysis of variance (ANOVA) in conjunction with a method for multiple comparisons, the Newman–Keuls test [6], to reduce the risk of random significance (
=0.05). The multiple significance tests performed in the Newman–Keuls method offer good protection against significant differences between imaging systems appearing "by chance". The statistical analysis was made using the software Statistica® Release 5.1 (StatSoft, Inc. Tulsa, OK).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
). Consequently, further analysis was based on the modified criteria unless otherwise stated.
|
HGolem and Kair,e for the different combinations of physical and technical parameters for the chest radiographs are shown in Table 3.
|
. Only 28% of radiographs were judged by the observers to fulfil all six modified image quality criteria. Criterion No. 10 showed the highest reproducibility. Criterion No. 9 had twice as many altered answers as criterion No. 10. The median value of the number of fulfilled criteria for each radiograph for all observers was used in this calculation. In fact, in no single case did all observers approve all criteria for a certain chest radiograph.
|
the ICS values are shown for six (excluding the "strict" observer) and seven observers, respectively. The ICSo values were calculated for all individual observers, and the highest and lowest of these values for six (excluding the "strict" observer) and seven observers are given in Figure 3. The minimum values for seven observers are all given by the same "strict" observer for all radiographic techniques.
|
|
. The techniques are arranged in order of decreasing values of ICS, and for each technique the corresponding values of VGAS are shown. The statistical analyses for ICS and VGAS are indicated by horizontal lines above and below the bars, respectively. Techniques can be separated from each other (p<0.05) only if they have no common horizontal lines. Techniques that cannot be statistically separated from each other are joined by a horizontal line. Techniques indicated by several horizontal lines cannot be separated from any of the techniques joined by these lines. In general, the agreement between the two types of scores regarding the relative quality of the different techniques is good. The two methods basically indicate the same techniques as best and worst. However, the ICS discriminates better between the different techniques, and the ICS is therefore used in the following analysis.
|
is a plot of ICS against HGolem. It can be seen that high dose techniques resulted in images with both many and few fulfilled criteria. Moreover, low dose techniques also resulted in images having both many and few fulfilled criteria. Thus, the number of fulfilled criteria cannot be used to unambiguously separate different techniques.
|
, the ICS and the radiation dose values HGolem for the 16 techniques have been averaged for the two states of each parameter (8 techniques per state). The table also shows the ratio between the two quantities (HGolem/ICS). It is clear that the different states of the parameters heavily influence the radiation dose as well as the image quality (see Discussion for further details).
|
shows the ICSc for each criterion at the two states of each of the four parameters. It can be seen that among the original CEC quality criteria, as well as among the modified ones, some criteria can be found that do not discriminate well between the different techniques. Examples are criteria No. 1 and No. 3 (original CEC quality criteria) and criteria No. 5 and No. 6 (modified quality criteria). Criteria that result in large variations in ICS are also seen, such as criteria No. 2 and No. 4 (original CEC quality criteria) and criteria No. 7 and No. 9 (modified quality criteria).
|
). For example, all techniques with density 1.8 had higher ICS values than those with density 1.3. The air gap technique gave, in general, lower patient radiation exposure values than the grid technique. The correlation between patient radiation exposure and image quality, as measured by the number of criteria fulfilled, was relatively poor. An example of good radiographic technique is 141 kVp, air gap, screen–film system with speed 320 and optical density 1.8. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Item (ii) is necessary for the optimization process to be successful in the radiological community and in fact rules out optimization strategies based on pure radiation physics alone. Item (iv) is the strongest of the prerequisites mentioned, since without an existing correlation between a given choice of available parameters of the radiographic process and the resulting "image quality", optimization cannot be performed with the tools available in the clinic.
As a result of CEC research projects and trials preceding the present one, the CEC presented "European guidelines on quality criteria for diagnostic radiographic images" in 1996 [3]. At the time of the trial these criteria existed in the form of a draft (the 1990 version) [2], but the differences between the versions were minor. The more or less ad hoc assumption of this study was that the CEC image quality criteria would meet the first three requirements mentioned above. The criteria certainly could be used to define image quality, they were relatively simple (but not, as will be discussed below, unambiguous), and they were based on clinical radiographic findings (anatomy) and therefore likely to be accepted. The possibility to quantify the criteria was not immediately seen and the predictability of the radiographic process using these criteria had to be tested.
Thus, Trial I of the present CEC project aimed at investigating (1) whether the CEC image quality criteria could be used for optimization of a radiographic process and (2) whether significant differences in image quality based on these criteria could be detected in a controlled project with well known physical and technical parameters.
The problem of quantifying the criteria was partly circumvented by the choice of study design of the first trial. The use of the criteria and the corresponding correlation with the physical and technical parameters was tested in two ways: in a simple "yes–no" procedure, where the criteria were evaluated in absolute terms, and by VGA where the criteria for a given image were compared with those of a reference image. A direct quantification of the criteria themselves was therefore not necessary. The succeeding statistical process could be regarded as accounting for the quantification, although this is not what one would generally mean by the word "quantification".
The properties of HGolem as compared with the International Commission on Radiological Protection (ICRP) definition of effective dose [7], and comparisons with other phantoms, are discussed in a paper by Zankl et al [5]. Specifically, the lung volume in the voxel model Golem is approximately 30% smaller than in Reference Man [8]. However, the lung density is correspondingly higher in the Golem model, leading to approximately the same lung mass. HGolem should therefore be similar to effective dose.
Any attempt to establish image quality without considering the "detectors" of the images, i.e. the radiologists, will be incomplete. The "detectors" of image quality in this trial were a group of European radiologists. Image formation and transfer of information to the observer of a diagnostic image are truly multifactorial processes, incorporating technical, physical, psychological and physiological elements. Consequently, the correlation between physical parameters and the diagnostic performance of radiographic exposures is difficult to establish. Predictivity of a diagnostic outcome from known physical and technical parameters, however, is a prerequisite for the prospects of establishing successful optimization of radiographic procedures. Therefore, efforts must be made to find methods and means of correlation for the prediction of procedures and to find a set of relevant clinical criteria to serve as indicators of image quality in the optimization process. The present project, and the trial reported here, is an attempt in this direction.
The image material
The total number of images and the number of images in each group of "image characteristics" were statistically optimized prior to the collection of images, thereby facilitating a reasonable statistical outcome of the trial. The use of volunteers and the consequent possibility to choose image characteristics without specific clinical demands also made the set-up of the chest part of the trial easier. Use of an air gap as a technique for scatter reduction was not mentioned in the CEC quality criteria, but it was tested here partly due to its potential for lower patient radiation exposure and results from previous studies [9].
Re-reading
The change in answers between reading sessions of an identical set of images was substantial. The reason for the large differences might be that the comparison contained the first of all the batches the observers were reading when they were inexperienced with the questions at issue. Although it was not tested, it could be assumed that the intrareader variability was considerably lower in the main part of the trial. It was also noted that the readers with better agreement between the readings were the ones who were more "positive" to the images (i.e. with more fulfilled criteria).
Radiation dose
As already indicated, the air kerma (kerma in free air) at the entrance surface (Kair,e) cannot beconsidered an adequate dose descriptor for patient dose in such trials, even when assessment of patient risk is not the predominant aspect in this context. This holds true mainly for the chest examinations. Indeed, the three groups with the highest mean patient doses in terms of HGolem (C3, C4 and C1; Table 3) and the group with the lowest (C14) also appear as the three highest and the lowest, respectively, in terms of Kair,e, and the ranking of the rest of the groups does not differ by more than three positions (C5) within the 16 groups. But while separation of the groups is statistically significant in terms of Kair,e, it often becomes insignificant in terms of HGolem because the values move together. The ratio between the highest dose (C3) and the second highest (C4) amounts to 1.51 in terms of Kair,e and to only 1.11 in terms of HGolem. In general, the increase of tube voltage from 102 kVp to 141 kVp reduces Kair,e by 33% on average while HGolem becomes only 8% lower. These findings confirm the assumption that the dose descriptor Kair,e indicates differences in patient dose that do not really exist with regard to patient risk. Beside the well known physical reasons, this effect is also caused by the definition of effective dose (from which HGolem is derived): the organs considered in the formation of effective dose preferably are positioned in the front half of the body and in the case of PA projections the influence of the higher X-ray penetrability becomes less efficient. Consequently, this difference between Kair,e and HGolem ratios is less pronounced with, for example, lumbar spine examinations [4]. As far as the statistical fluctuations of the dose values permit such considerations (see below), and after correcting Kair,e for patient thickness, ranking and ratios between the various patient groups could be expressed equally well in terms of Kair,e or HGolem.
Relationship between image-producing procedures and patient dose
In the case of chest examinations in this study, the image-producing procedures are clearly mirrored by the HGolem values. The use of an air gapinstead of a grid (C9/C1, C10/C2,...C16/C8; Table 3) reduces HGolem by on average 52% (49–57%). The images with an optical density of 1.3 needed 32% (24–41%) less dose than those with optical density 1.8 (C1/C3, C2/C4,...C14/C16). The shift from speed class 160 to speed class 320 (C5/C1, C6/C2,...C16/C12) reduces patient dose by 37% (32–42%) for the 102 kVp spectra and by 45% (41–48%) for the 141 kVp spectra. Speed class values 160 and 320 are of nominal character. The real speed of the two screen–film systems was 213 and 389, respectively, as measured under the conditions of ISO standard 9236 behind a phantom for a 120 kV spectrum [10]. The patient dose ratio between speed classes 320 and 160 should therefore amount to 0.55, and this was exactly found for the 141 kVp examinations. The moderate deviation for the 102 kVp spectra is not surprising and is possibly due to adifference in screen–film response between the 102 kVp spectrum and the 120 kVp spectrum used in the speed measurement.
A purely calculatory estimation of the effect to be expected when tube voltage is increased from 102 kVp to 141 kVp was performed by normalizing HGolem to an arbitrary constant dose value behind the patient (different beam absorption in patient couch and grid and different screen–film response is not considered in this approach). This procedure led to a patient dose reduction of 6% for the grid and 10% for the air gap technique when tube voltage was increased from 102 kVp to141 kVp. On average, the ratio of doses for different tube voltages found in the trial (C2/C1, C4/C3,...C16/C15) amount to a value of 0.92, which is in excellent agreement with the above values, but the single values vary from 0.78 (C16/C15) to 1.04 (C10/C9). Facing the insignificance of the effect and the statistical fluctuations of the underlying measured air kerma values, the result can be regarded as satisfactory.
The relation between radiation dose and image quality
As expected, the differences between patient radiation exposure for the various image characteristics were large. The correlation between the number of fulfilled criteria and patient radiation exposure was found to be small (Figure 5). This, however, leads to a possibility of optimizing the examination, since techniques with both low radiation dose and high percentage of fulfilled criteria can be found.
Many interesting observations can be made from Table 4. The tube voltage appears to have a small influence on the ICS as well as on the radiation dose, as already mentioned. Using 141 kVp leads to a smaller HGolem/ICS compared with the use of 102 kVp, however, since the ICS is higher by a small amount, whereas the radiation dose is reduced by a small factor. For the method of scatter reduction, a much larger difference can be seen. The two methods received almost identical ICS values, although the radiation dose differs by a factor of two, to the advantage of the air gap technique. The same discussion can be made for the speed class, to the advantage of the faster film–screen combination. In the case of theoptical density, the HGolem/ICS is almost the same, but the 1.8 optical density technique receives a much higher ICS at the expense of a much higher dose.
Figure 5 clearly demonstrates that, in most cases, higher patient doses cannot be justified by the desire for higher image quality and that there exists a large potential to reduce patient dose without any loss of image quality. Consequently, it would be better to choose a technique with lower dose than to choose the technique with the highest number of fulfilled criteria. The combination of speed class 320, optical density 1.8, air gap and tube voltage 141 kVp resulted in both a high ICS and a low HGolem and would be a good compromise between image quality and radiation dose to the patient.
Methods for evaluation
From Figure 4 it can be seen that the two methods used here for evaluation, the ICS and the VGAS, indicated the same techniques as the best and worst ones. All chest images were of high technical and "radiographical" quality, which was ensured by the use of volunteers, the use of one dedicated chest stand as well as the controlled set of physical and technical parameters. All images were taken by two specially trained radiographers. However, only 28% of the images were judged by the observers (median score) to fulfil all six quality modified criteria. This could in part be explained by the fact that not all physical and technical parameters were chosen with the intent to produce an optimized chest image. Despite this, it is questionable whether the quality criteria alone could be used to evaluate the usefulness of a specific combination of technical and physical parameters in clinical practice.
The use of the image criteria as a single tool for judging image quality is open to discussion. The importance of the observer is significant, as shown by the results of the present study. All the images included in the trial were judged to be clinically useful, but less than one-third of the images fulfilled all image criteria. Also, taking patient radiation exposure into consideration, the number of fulfilled criteria (the ICS) could not be used to unambiguously separate different techniques from each other, since the correlation between patient radiation exposure and ICS was poor. Thus, the image criteria alone cannot be used when judging whether the image quality is acceptable for clinical use. This is not surprising, since, in general, image quality is hardly ever well correlated with patient radiation exposure. The result of an optimization process is also influenced by the images included in the test and by the observers involved. The overall impression of the images when setting ones threshold could be important as well as the quality of the images the observers are used to. This means that the results when using the image criteria cannot be compared if different observers are used for different images. Training in using these image criteria is vital. Reference radiographs would be useful for the observer. Also, the wording used in the image criteria is very important for consistent interpretation. The revision work performed so far has definitely resulted in criteria with much of the semantic ambiguities eliminated.
The discriminating capabilities of the criteria were tested in the present trial by analysing the statistical significance of the differences in the two quality scores (ICS and VGAS) found between the techniques used. Indeed, significant differences were found between groups of techniques. Although the differences between the techniques were not very subtle in this case, it is promising for the possibility of predicting the outcome of specific parameter settings in radiography that such differences are found in a controlled project. The discriminating capabilities have been evaluated further in the present CEC project: the modified criteria have been tested on digitized versions of the same image material as in the present study, manipulated in terms of noise, resolution and contrast [11, 12].
Some of the quality criteria did not discriminate well between the different techniques (Figure 6). The use of these specific criteria for future studies should therefore be questioned, since it leads to a higher workload on the observer and to a loss in statistical separation power. However, since all criteria are chosen based on their clinical relevance, there is a danger in omitting some criteria since a situation may occur in which their use would have led to the finding of an unsatisfactory technique.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Regarding the first item, use of the present image criteria was not unambiguous and, when optimizing radiographic techniques, the results could be difficult to interpret owing to the influence of specific observers. In a trial of this kind, the quality of the image criteria is essential for the results. In addition to the question of the usefulness of the criteria in an optimization process, it should also be clarified whether the image criteria work as intended as a tool for evaluating good practice.
Discussion of the trial results shows that it is possible to design trials under routine clinical conditions that provide the desired information in a reliable way, even when they are conducted with realistic patient collectives varying greatly in size, weight and thickness. This presupposes that all physical, technical and image-related parameters are strictly audited and that the influence of disturbing contributing factors is either avoided or considered in the trial evaluation. If these pre-conditions are fulfilled, even a finding that groups of examinations performed under different image-producing procedures cannot be separated in a significant way owing to statistical fluctuations caused by patient variations should not be seen as a failure or insufficiency of the trial. Rather it is an indication that the underlying differences are of marginal importance with regard to either patient dose or radiological image content and therefore need no further consideration.
Regarding the technical parameters, use of an air gap technique is recommended compared with the use of a grid, since it reduces the dose to the patient while maintaining the same image quality. The same also holds true for the speed of the screen–film system, where speed class 320 results in the same image quality as speed class 160, but the absorbed dose to the patient is halved with 320. The influence of the tube voltage is minor. However, 141 kVp resulted in both a slightly lower radiation dose to the patient and a somewhat higher image quality. For the optical density, the conclusion is that the use of optical density 1.8 leads to a higher image quality than the use of 1.3, but at the cost of a higher absorbed dose to the patient.
In summary, the criteria can be used to highlight optimum radiographic technique in terms of image quality and patient dose, although not unambiguously. A recommendation for good radiographic technique, based on a compromise between image quality and a risk-related radiation dose to the patient, is to use 141 kVp, an air gap, a screen–film system with speed 320 and optical density 1.8.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication July 31, 2000. Revision received July 13, 2001. Accepted for publication August 13, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M Bath and L G Mansson Visual grading characteristics (VGC) analysis: a non-parametric rank-invariant statistical method for image quality evaluation Br. J. Radiol., March 1, 2007; 80(951): 169 - 176. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G McVey The effect of phantom type, beam quality, field size and field position on X-ray scattering simulated using Monte Carlo techniques Br. J. Radiol., February 1, 2006; 79(938): 130 - 141. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Borjesson, M. Hakansson, M. Bath, S. Kheddache, S. Svensson, A. Tingberg, A. Grahn, M. Ruschin, B. Hemdal, S. Mattsson, et al. A software tool for increased efficiency in observer performance studies in radiology Radiat Prot Dosimetry, May 17, 2005; 114(1-3): 45 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tingberg, M. Bath, M. Hakansson, J. Medin, J. Besjakov, M. Sandborg, G. Alm-Carlsson, S. Mattsson, and L. G. Mansson Evaluation of image quality of lumbar spine images: a comparison between FFE and VGA Radiat Prot Dosimetry, May 17, 2005; 114(1-3): 53 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tingberg, F. Eriksson, J. Medin, J. Besjakov, M. Bath, M. Hakansson, M. Sandborg, A. Almen, B. Lanhede, G. Alm-Carlsson, et al. Inter-observer variation in masked and unmasked images for quality evaluation of clinical radiographs Radiat Prot Dosimetry, May 17, 2005; 114(1-3): 62 - 68. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tingberg and D. Sjostrom Optimisation of image plate radiography with respect to tube voltage Radiat Prot Dosimetry, May 17, 2005; 114(1-3): 286 - 293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Grahn, B. Hemdal, I. Andersson, M. Ruschin, A. Thilander-Klang, S. Borjesson, A. Tingberg, S. Mattsson, M. Hakansson, M. Bath, et al. Clinical evaluation of a new set of image quality criteria for mammography Radiat Prot Dosimetry, May 17, 2005; 114(1-3): 389 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Tingberg, C Herrmann, B Lanhede, A Almen, M Sandborg, G McVey, S Mattsson, W Panzer, J Besjakov, L G Mansson, et al. Influence of the characteristic curve on the clinical image quality of lumbar spine and chest radiographs Br. J. Radiol., March 1, 2004; 77(915): 204 - 215. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
) and seven observers (
).
