British Journal of Radiology 74 (2001),615-620 © 2001 The British Institute of Radiology
Procedure for quantitatively assessing automatic exposure control in mammography: a study of the GE Senographe 600 TS
S Meeson, PhD1,,
K C Young, PhD1,
P B Hollaway, MSc2 and
M G Wallis, FRCR3
1National Co-ordinating Centre for the Physics of Mammography and 2Regional Radiation Protection Service, Department of Medical Physics, St Luke's Wing, Royal Surrey County Hospital, Guildford GU2 7XX, and 3Warwickshire, Solihull & Coventry Breast Screening Centre, Coventry and Warwickshire Hospital, Stoney Stanton Road, Coventry CV1 4FH, UK
 |
Abstract
|
|---|
The correct operation of a mammography system's automatic exposure control system (AEC) is essential if mammograms are to be produced with a suitable film exposure. A methodology has been developed that allows the performance of the AEC to be assessed quantitatively using clinical films. By digitizing mammograms, the mean optical density (OD) in the main breast region and in a region of interest corresponding to the position of the AEC detector are evaluated for each film, together with the area of the main breast. Using these data it is possible to determine the relationships between the mean OD, breast size and AEC detector position. The performance of the AEC on a GE Senographe 600 TS system was investigated. The study found that there is a tendency to underexpose smaller breasts, i.e. with an area less than approximately 4000 mm2. This is equivalent to a compressed tissue width of approximately 60–80 mm. The difference in mean OD between the mammograms of small and large breasts was up to 0.7 OD. Provided the sensitive area of the AEC detector is known, this method of assessing AEC performance can be used with any mammography system.
|
Introduction
|
|---|
The production of consistently high quality mammograms is one of the major objectives of the extensive quality system implemented by the UK National Health Service Breast Screening Programme (NHSBSP). To achieve this, film–screen mammography equipment and imaging procedures need to be optimized. This includes ensuring that mammograms are suitably exposed. In film–screen mammography, the exposure is normally controlled by an automatic exposure control system (AEC). An AEC monitors the X-rays transmitted through the breast. The exposure is terminated when the radiation received by a detector reaches a predetermined level corresponding to the desired optical density (OD) on the film [1]. With most mammography systems, the position of the AEC detector can be adjusted between two or more positions. These positions are selected at the discretion of the operator, usually based on the size of the breast. Correct operation of the AEC is essential if the mean OD in the main breast region of each mammogram is to achieve the desired value. As part of quality assurance (QA), AECs are checked using test blocks of polymethyl methacrylate (PMMA) following standard procedures [1, 2]. However, it is not well established how measurements using blocks of PMMA relate to the film densities in clinical mammograms. A methodology is presented that allows the performance of an AEC to be assessed quantitatively using clinical films. The method has been used to study the operation of the AEC on a GE Senographe 600 TS mammography X-ray system (GE Medical Systems Europe, Paris, France).
During an annual survey of a GE 600 TS X-ray unit, the densities of test films were found to have been strongly affected by the size of the PMMA test blocks used routinely. This had led to miscalibration of the AEC and wrongly exposed clinical films. This caused us to investigate how the AEC performance changed for different areas of test blocks and also for different sizes of breast.
|
Methods
|
|---|
Tests using different areas of PMMA
During routine QA visits by physicists to four sites in the South East region with GE Senographe 600 TS systems, a series of additional exposures was made to test the effect of the test block area. Each system was tested using a 40 mm thick block of PMMA with the AEC detector in the inner position, closest to the chest wall. Starting from the chest wall edge of the breast support table, the block was placed covering widths of 20, 40, ..., 180 mm of the support table (Figure 1
). Using the current AEC density setting, an exposure was made for each block position. The post-exposure mAs was recorded and the relationship between the mAs selected and the width of PMMA covering the support table was determined for each system. Both rectangular 190 x 240 mm2 and semi-circular 120 x 240 mm2 shaped test objects were used to assess whether the shape of the test object was important. During the last QA visit, test films were created for each test block position. The X-ray set was operated under automatic exposure control. A kilovoltage of 28 kV was used for the films. The AEC detector chamber was positioned at the chest wall. An Imation HCM film–screen combination was used and the films were processed using an Imation Trimatic processor operating on extended development cycle.
View larger version (51K):
[in this window]
[in a new window]
|
Figure 1. Plan of the breast support table (not to scale) showing where the test blocks were positioned for the exposure measurements. The shaded area corresponds to the area of the support table that was covered when the rectangular test object was placed 160 mm from the chest wall edge of the support table. The positions of the test object when placed 40 mm, 80 mm and 120 mm from the chest wall edge of the support table are also indicated. When semi-circular blocks were used, the distance from the chest wall was measured to the furthest point on the semi-circle.
|
|
Analysis of clinical films
To test the performance of the AEC using clinical films, a sample of screening films was digitized, calibrated and analysed. The clinical film sample comprised 122 mammograms (89 mediolateral oblique views and 33 craniocaudal views) from 45 women attending a single day of screening on a mobile unit in the West Midlands region. The ages of the women ranged from 49 years to 64 years, with a mean±standard deviation (SD) of 55.7±4.3 years. The X-ray set used was a GE Senographe 600 TS operated under automatic exposure control. A kilovoltage of 28 kV was used for the majority of the films. However, at the radiographer's discretion, 30 kV was selected for a small number of women with large breasts. The routine practice was to have the AEC chamber positioned at the chest wall for the quality control films (i.e. using blocks of PMMA) and to move the chamber depending on breast size for the screening films. For each exposure, the radiographer recorded the position of the AEC detector. Kodak min-R 2000 film was used with a Kodak min-R 2000 screen. The films were processed using a Kodak M35M processor using Kodak RP X-Omat EX chemicals operating at a developer temperature of 35 °C on a standard cycle (150 s dry-to-dry).
The methods of film digitization and calibration have been described previously [3]. Films were digitized using a Lumiscan 150 HR laser scanner (Lumisys, Sunnyvale, CA) at 210 µm resolution. The scanner's detected signal was digitized into image pixel values. Grey scale step wedge films (using Kodak min-R 2000) were digitized along with the clinical films to facilitate the conversion from image pixel value to optical density units. An image processing software package, Aphelion (Amerinex Applied Imaging, Inc., Northampton, MA), was used to create and store a pixel value to OD calibration. This was then used to produce images with pixels in OD units.
Digitized images of the mammograms were first analysed using three mutually exclusive regions of interest (ROIs) representing the pectoral muscle (oblique films only), main breast and skin edge region [3]. A semi-automatic means of selecting the ROIs was used [4]. Image analysis performed on each digitized mammogram included measuring the area and the mean OD in each ROI. The GE 600 TS system has three possible AEC detector positions. The locations and dimensions of the sensitive areas of the AEC detector, in the plane of the film, were established using technical data provided by GE. The positions of the sensitive areas of the AEC detector were mapped onto a film and digitized along with the clinical films at the same resolution. Three ROIs could then be defined corresponding to the inner (or chest wall),middle and outer AEC detector positions. Figure 2 shows images of digitized mammograms with the ROIs corresponding to the AEC detector positions overlaid. The digitized images of the mammograms were analysed to determine the maximum, minimum and mean OD in each of these three ROIs. Finally, relationships between the mean OD in the main breast ROI, the mean OD in the ROI corresponding to the AEC detector position and breast size were examined.
View larger version (121K):
[in this window]
[in a new window]
|
Figure 2. The three possible regions of interest corresponding to the automatic exposure control detector positions are shown on the mammogram images of women with predominantly adipose (a) and dense glandular (b) breast tissues.
|
|
Analysis of films of PMMA
Using the above protocol for the analysis of clinical films, each film of PMMA was digitized, calibrated and the mean OD in the ROI corresponding to the AEC detector position calculated. For these films, the area of the main breast was taken to be the area of the image of the semi-circular block of PMMA. The relationship between the mean OD in the ROI corresponding to the AEC detector position and the area of the main breast was established.
|
Results
|
|---|
Tests using different areas of PMMA
Figure 3 shows the relationship between the width of the test object covering the support table and the post-exposure mAs, selected by the AEC. For the four systems examined, the exposure measured in mAs reached a constant or plateau value at widths of PMMA greater than or equal to 100 mm. The selected mAs data were normalized for each imaging system. At widths of PMMA greater than 20 mm the entire AEC detector was covered. As the test object width increased from 20 mm to 60–80 mm, the normalized mAs increased rapidly with increasing coverage of the breast support table. The same behaviour was observed with both rectangular (X-ray sets 2, 3 and 4) and semi-circular (X-ray set 1) shaped test objects.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. Relationship between the width of the test object and the normalized exposure for four different GE 600 TS systems.
|
|
Analysis of clinical films
The relationship between the mean OD in the main breast and the area of the main breast ROI is shown in Figure 4. As the area of the main breast ROI increased, the position of the AEC detector selected by the radiographer changed from the inner position to the middle position and finally to the outer position. The range of mean ODs and the overall mean OD (±SD) in the main breast ROIs for each AEC detector position are listed in Table 1. The range of mean OD was widest for the AEC detector in the inner position, and narrowest for the AEC detector in the outer position. When the AEC detector was in the inner position the mean OD in the main breast ROI was occasionally relatively low (i.e. <1.5). When the AEC detector was in the middle position there was a fairly wide spread of densities around a mean value of 1.86±0.13 OD. The AEC produced the most reproducible mean OD, with the smallest SD, for large breasts with the AEC detector in the outer position. The overall mean OD in the main breast was lowest for breast areas of less than 2000 mm2, 1.63±0.06 (2 standard errors in the mean), rising to 1.86±0.04 for 2000–4000 mm2 and 1.84±0.06 for breast areas greater than 4000 mm2.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 4. Relationship between the mean optical density (OD) in the main breast region of interest (ROI) and the area of the main breast ROI. The position of the automatic exposure control detector used clinically is indicated by the data point symbol.
|
|
View this table:
[in this window]
[in a new window]
|
Table 1. Mean optical densities (ODs) in the main breast regions of interest of clinical films, for each automatic exposure control (AEC) detector position
|
|
Figure 5 shows the mean ODs in the ROIs corresponding to the AEC detector positions rather than in the whole main breast area. By using just the ROI corresponding to the AEC detector position, the density variations due to different breast tissues within the breast were reduced. The relationships observed in Figure 4
are more noticeable in Figure 5, where the mean ODs in the ROIs corresponding to the AEC detector positions were plotted against the areas of the main breast ROIs. The clinical films were relatively underexposed for small breasts. For small breasts, the mean ODs in the ROIs corresponding to the AEC detector positions were below the mean OD value achieved for the larger breasts. The mean ODs are listed in Table 2. The overall mean ODs in the ROIs corresponding to the AEC detector positions were all slightly lower than for the main breast ROIs. Comparing both Figures 4 and 5 and the SD data in Tables 1 and 2, there was less spread in the densities for the ROIs corresponding to the AEC detector positions than for the main breast ROIs. In Figures 4 and 5 the mean OD in the ROIs reached a plateau for breast areas greater than approximately 4000 mm2.
View larger version (21K):
[in this window]
[in a new window]
|
Figure 5. Relationship between the mean optical density (OD) in the region of interest (ROI) corresponding to the automatic exposure control (AEC) detector position and the area of the main breast ROI. The position of the AEC detector used clinically is indicated by the data point symbol. The corresponding ODs for a semi-circular block of PMMA are shown as a solid curve. The PMMA data are from a different X-ray system (X-ray set 1) and are shown together with the clinical data to demonstrate the similarity in the shape of the relationship between the mean OD in the ROI corresponding to the AEC detector position and the area of the main breast.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Mean optical densities (ODs) in the regions of interest corresponding to the automatic exposure control (AEC) detector positions in clinical films
|
|
Analysis of films of PMMA
The relationship between the mean OD in the ROI corresponding to the inner AEC detector position (X-ray set 1) and the area of the semi-circular PMMA test object is shown in Figure 5 as a solid line, together with the clinical film data. For both PMMA and clinical films the mean OD in the ROI corresponding to the inner AEC detector position increased with increasing breast size. The mean OD in the films of PMMA reached a plateau for areas greater than approximately 4000 mm2.
|
Discussion
|
|---|
The mean ODs in the ROIs corresponding to the AEC detector positions were lower and less variable than the mean ODs in the main breast ROIs. This may be explained as a consequence of the mean ODs in the main breast ROIs being evaluated over a larger area and containing more widely varying breast tissue types.
It can be seen in Figures 4 and 5 that the clinical films were relatively underexposed for small breasts. For these small breasts, the mean ODs in the main breast and in the ROIs corresponding to the AEC detector positions were below the mean OD value achieved for the larger breasts. This trend was also observed in Figure 3 where the mAs was relatively low for small test object sizes, but increased with increasing test object area until a plateau mAs was achieved. The data for X-ray sets 3 and 4 did not exactly match the data for sets 1 and 2 for small widths of PMMA. This is likely to be the result of differences in the exact location of the AEC detector within the breast support tables. Perhaps this is understandable considering the age of the equipment involved. The mean ODs in the ROI ofboth PMMA and clinical films were low for small breast areas with the AEC detector in the innerposition. This study has shown that GE Senographe 600 TS X-ray sets underexpose small areas (and widths) of a 40 mm thick block of PMMA. The underexposure of smaller breasts was also found clinically with this design of X-ray set. While breast thickness can be expected to vary with breast size, the similarity of the two sets of results suggests that breast area is the dominant factor in the underexposure of small breasts. The effect of varying the thickness of PMMA blocks was tested, along with width across the breast support table, and while thickness had little effect on a well calibrated AEC, underexposure was always found where the width of PMMA was small. Clinically, it may be possible to compensate for such underexposure of small breasts by increasing the density setting for this group of women. In this study, using the GE Senographe 600 TS with the film–screen processing described, increasing the density to +1 or +2 may be sufficient to compensate for the underexposure.
It is likely that the underexposure of small breasts is the result of the large number of direct X-rays (i.e. that have not passed through the breast) entering the breast support table and scattering towards the AEC detector leading to premature termination of the exposure. This does not happen for larger breasts, which cover a greater area of the support table, since the X-ray beam cannot directly reach the area of the support table adjacent to the AEC detector without attenuation.
Standard QA measurements are made using test blocks of PMMA. However, it is not well established how measurements using these blocks relate to the film densities in clinical mammograms. Previous studies have found that the mean ODs in the main breast region were higher than the density measurements made on routine QA films of PMMA [3, 5]. In this study, clinical films for average and large sized breasts had mean ODs in the main breast region that were again above the OD measured on QA films of a 40 mm thickness of PMMA, by about 0.2 OD. (As discussed, the mean ODs were lower for small area breasts.) It is assumed that this effect is caused by the heterogeneous nature of breast tissue compared with test blocks.
The correct operation of the AEC is essential if consistent ODs for different breast types are to be achieved. The positioning of the AEC detector can also greatly influence the mean OD in the main breast region of a mammogram because of the variability of the distribution of breast tissues. Images of digitized mammograms in Figure 2 demonstrate this. Figure 2a shows that with the AEC detector in the inner position, the tissue above the AEC detector was mainly pectoral muscle. As a result a relatively large exposure resulted and a relatively high mean film OD in the main breast was produced. Figure 2b further shows how unknown breast physiology of women being screened can affect the film exposure. With the AEC detector again in the inner position, the tissue above the detector was predominantly adipose, whereas there was mostly glandular tissue elsewhere in the breast resulting in a relatively low mean OD in the main breast ROI overall. Retrospectively it can be judged that if the AEC detector had been in the middle or the outer position a better exposure of the dense breast tissues would have been achieved.
|
Conclusions
|
|---|
A quantitative method of assessing AEC performance has been developed and demonstrated. This has shown that with GE Senographe 600 TS X-ray sets there is a tendency to underexpose smaller breasts, resulting in relatively low mean optical densities. Provided the position of the sensitive area of the AEC detector is known, this method of assessing AEC performance can be used with any mammography system. Evidence suggests that AEC systems tend to produce films with lower mean optical densities for homogeneous blocks of PMMA than for inhomogeneous breast tissue. On the basis of this study, it is recommended that consideration should be given to increasing the density setting on GE 600 TS X-ray sets when imaging women with small breasts when the AEC detector is in the inner position. The blocks of PMMA used routinely to test AEC performance should have a width (from chest wall towards the nipple) of at least 100 mm and preferably 150 mm.
|
Footnotes
|
|---|
The work of the National Co-ordinating Centre for the Physics of Mammography is funded by the National Co-ordinating Office of the NHS Breast Screening Programme. 
Received for publication August 9, 2000.
Accepted for publication April 4, 2001.
|
References
|
|---|
-
Law J, Dance DR, Faulkner K, Fitzgerald MC, Ramsdale ML, Robinson A. The commissioning and routine testing of mammographic X-ray systems, IPSM Report No. 59 (2nd edn). York: Institute of Physical Sciences in Medicine, 1994.
-
Institute of Physics and Engineering in Medicine. Recommended standards for the routine performance testing of diagnostic X-ray imaging systems, IPEM Report No. 77 (1st edn). York: IPEM, 1997.
-
Meeson S, Young KC, Ramsdale ML, Wallis MG, Cooke J. Analysis of optical density and contrast in mammograms. Br J Radiol 1999;72:670–7.[Abstract]
-
Meeson S, Young KC, Rust R, Ramsdale ML, Wallis MG, Cooke J. Implications of using high contrast mammography X-ray film–screen combinations. Br J Radiol (in press).
-
Young KC, King S, Meeson S, Ramsdale ML, Cooke J, Wallis MG. Auditing the image quality of mammograms. Radiat Prot Dosim 2000;90:279–82.[Abstract]
Top
Abstract
Introduction
