TECHNICAL EVALUATION OF THE KONICA MINOLTA REGIUS 190 CR MAMMOGRAPHY SYSTEM AND THREE TYPES OF IMAGE PLATE
NHSBSP Equipment Report 0806 December 2008 KC Young, JM Oduko, O Gundogdu and A Alsager National Coordinating Centre for the Physics of Mammography
Enquiries Enquiries about this report should be addressed to: Professor KC Young National Coordinating Centre for the Physics of Mammography Medical Physics Department Royal Surrey County Hospital Guildford GU2 7XX Tel: Fax: Email: ken.young@nhs.net Published by NHS Cancer Screening Programmes Fulwood House Old Fulwood Road Sheffield S10 3TH Tel: 1060 Fax: 1089 Email: info@cancerscreening.nhs.uk Website: www.cancerscreening.nhs.uk NHS Cancer Screening Programmes 2008 The contents of this document may be copied for use by staff working in the public sector but may not be copied for any other purpose without prior permission from the NHS Cancer Screening Programmes. The report is available in PDF format on the NHS Cancer Screening Programmes website.
Typeset by Prepress Projects Ltd, Perth (www.prepress-projects.co.uk) Printed by Henry Ling Limited
Technical Evaluation of the Konica Minolta Regius 190 CR Mammography System

Contents

1. 1.1 1.2 2. 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 3. 3.1 3.2 3.3 3.4 3.5 3.6 3.7 4. 5.
ACKNOWLEDGEMENTS INTRODUCTION Testing procedures and performance standards for digital mammography Objectives METHODS System tested Detector response Dose measurement Contrast to noise ratio Noise analysis Image quality measurements Physical measurements of the detector performance Image retention RESULTS Detector response Dose and CNR under AEC control Noise measurements Image quality measurements Image retention Comparison with other systems Detector performance DISCUSSION CONCLUSIONS REFERENCES

NHSBSP December 2008

ACKNOWLEDGEMENTS
The authors are grateful to the staff at Radiology Department of Royal Surrey County Hospital, Guildford, for their help in evaluating the unit at their site.

1. INTRODUCTION

1.1 Testing procedures and performance standards for digital mammography
This report is one of a series evaluating commercially available digital mammography systems on behalf of the NHS Breast Screening Programme (NHSBSP). The testing methods and standards applied are mainly derived from NHSBSP Equipment Report 0604.1 This is referred to in this document as the NHSBSP protocol and it has the same image quality and dose standards as those provided in the European protocol.2 The European protocol was followed where there is a more detailed performance standard, eg for the automatic exposure control (AEC) system. 1.2 Objectives
The purpose of these tests was to determine whether this system meets the main standards in the NHSBSP and European protocols and to provide performance data for comparison against other manufacturers products. Clinical evaluations are published separately by the NHSBSP where systems meet the minimum standards in the NHSBSP protocol. A final decision on the suitability of systems for use in the NHSBSP depends on a review of both the technical and clinical evaluations.

2. METHODS

2.1 System tested
The tests were conducted at the Radiology Department of Royal Surrey County Hospital, Guildford, using the Konica Minolta Regius 190 computerised radiography (CR) reader shown in Figure 1. The reader produces images with a pixel size of 43.75m. Three different types of mammography CR plate were evaluated and compared (Table 1). The RP-6M plate is the design previously available for use with the system. The RP-7M is an improved version of the RP-6M. The CP-1M is a new type of plate with a phosphor that has a needle crystal technology designed to improve the sharpness of the images by reducing light spread. Such technology is available for general radiology but has not previously been used in a commercial mammography CR system. The exposures were made using a GE DMR mammography x-ray set. The choice of kV, target material and filter material and mean glandular dose (MGD) were those recommended by Konica Minolta for use with the RP-7M plates (Table 2). The AEC was not adjusted as no patients were to be exposed and the factors were set manually with the aim of achieving the dose levels given in Table 2. As a result, the three plate types were given identical exposures during the testing. Although Konica Minolta provided these recommended dose levels, they pointed out that they believe it may be possible to employ lower dose levels due to the improved detector efficiency of the two newer types of plate and depending on the results of the tests in this report. 2.2 Detector response

The response of each type of CR cassette was measured broadly as described in the NHSBSP protocol. A phantom of Plexiglas (polymethylmethacrylate; PMMA) with a total thickness of 45mm was positioned at the tube exit port and exposed using the three target/filter combinations available (Mo/Mo, Mo/Rh and Rh/ Rh) at tube voltages spanning the range used clinically (25, 28, 31, 34 and 37 kV). A cassette was placed on
Figure 1 The Konica Minolta Regius 190 CR reader.
Table 1 Image plates tested in this report Manufacturer Konica Minolta Konica Minolta Image plate name RP-6M RP-7M Material BaFI BaFI Comment RP-6M is the generation previously available RP-7M is the seventh generation of coating type plate and was specially designed to meet the criteria specified in French regulations CP-1M is a new plate design using needle crystal technology to achieve higher sharpness and higher detective quantum efficiency (DQE)

Konica Minolta

Table 2 Exposure parameters recommended by Konica Minolta with RP-7M plates Thickness of PMMA (mm) 70 Equivalent breast thickness (mm) 90 Tube voltage (kV) Target 29 Mo Mo Mo Mo Mo Mo Mo Filter Mo Mo Mo Mo Mo Rh Rh Mean glandular dose (mGy) 0.85 1.28 1.70 2.13 2.70 4.28 6.18
top of the bucky, an ion chamber was positioned on the top surface of the cassette, and the entrance surface air kerma was measured for a single tube current-time product for each tube voltage and target/filter combination tested. In addition, the entrance air kerma was measured for a wide range of tube current-time product values at 28kV Mo/Mo. The readings were corrected to the surface of the cassette using the inverse square law. No correction was made for attenuation by the top cover of the cassette. The images were saved as unprocessed files and transferred to another computer for analysis. Multiple 2.5 mm square regions of interest (ROIs) were positioned on the midline and 6cm from the chest wall edge of each image. These small ROIs were used to minimise the influence of the heel effect on the measurement of standard deviation.3 The average pixel value and the standard deviation of pixel values within that region were measured. The relationship between average pixel values and the detector entrance surface air kerma was determined. This was used to linearise all pixel values and their standard deviations against the entrance surface air kerma. 2.3 Dose measurement

The relative CNR was then calculated according to equation 2 and compared with the limiting values provided for relative CNR in the European protocol and shown in Table 3. The minimum CNR required to meet this criterion was then calculated. Relative CNR = CNRmeasured/CNRlimiting value (2)

Noise analysis

The 45mm thick block of PMMA with aluminium square used for the measurement of CNR was exposed using manually selected tube loading across the range available. The kV target/filter combination used was that selected by the AEC in standard dose mode. The compression paddle was in place. The same regions of interest used in the CNR measurement were applied to the corresponding unprocessed images. All pixel values and standard deviations measured were linearised with respect to entrance air kerma at the surface of the detector (ie the CR cassette) as described in the NHSBSP protocol. After linearisation, the average standard deviations of the pixel values in the background ROI for each image were used to investigate the relationship between dose to the detector and image noise. It was assumed that this noise comprises three components electronic noise, structural noise and quantum noise with the relationship shown in equation 3. This method of analysis has been described previously.4
Table 3 Limiting values for relative CNR Thickness of PMMA (mm) 70 Equivalent breast thickness (mm) 90 Limiting values for relative CNR (%) in European protocol >115 >110 >105 >103 >100 >95 >90
2 p = ke2 + kq p + ks2 p 2
where p is the standard deviation in pixel values within an ROI with a uniform exposure and a mean linearised pixel value p, and ke, kq and ks are the coefficients determining the amount of electronic, quantum and structural noise in a pixel with a value p. For simplicity, the noise is generally presented here as relative noise defined as in equation 2. Relative noise =
The variation in relative noise with mean pixel value was evaluated and fitted using equation 3, and non-linear regression was used to determine the best fit for the constants and their asymptotic confidence limits (using Graphpad Prism Version 5 for Windows, Graphpad software, San Diego, California, USA; www.graphpad. com.). This established whether the experimental measurements of the noise fitted this equation and the relative proportions of the different noise components. In fact, the relationship between noise and pixel values has been found empirically to be approximated by a simple power relationship, as shown in equation 5.

= kt p n

where kt is a constant. If the noise were purely quantum noise, the value of n would be 0.5. However, the presence of electronic and structural noise means that n can be slightly higher or lower than 0.5. 2.6 Image quality measurements
Contrast detail measurements were made using the CDMAM phantom (version 3.4, UMC St. Radboud, Nijmegen University, Netherlands). The phantom was positioned with a 20mm thickness of PMMA above and below, to give a total attenuation approximately equivalent to 50mm of PMMA or 60mm thickness of typical breast tissue. This arrangement was imaged using the x-ray sets AEC in standard mode with small rigid expanded polystyrene spacers in place to create a total thickness of 60mm. This procedure was repeated with small adjustments to the phantom position to obtain a representative sample of 16 images at this dose level. Unprocessed images were transferred to disk for subsequent analysis off-site. Further images (sets of 16) of the test phantom were then obtained at other dose levels by manually selecting higher and lower mAs values with the same beam quality as selected under AEC control. An automatic method of reading the CDMAM images was used.57 The threshold gold thickness for a typical human observer was predicted using equation 6. TCpredicted = rTCauto (6)
where TCpredicted is the predicted threshold contrast for a typical observer and TCauto is the threshold contrast measured using an automated procedure with CDMAM images. Contrasts were calculated from gold thickness for a nominal tube voltage of 28kV and a Mo/Mo target filter combination as described in the European protocol. r is the average ratio between human and automatic threshold contrast determined experimentally with the values shown in Table 4.5
Table 4Values of r used to predict threshold contrast Diameter of gold disc (mm) 0.08 0.10 0.13 0.16 0.20 0.25 0.31 0.40 0.50 0.63 0.80 1.00 Average ratio of human to automatically measured threshold contrast (r) 1.40 1.50 1.60 1.68 1.75 1.82 1.88 1.94 1.98 2.01 2.06 2.11
The main advantage of automatic reading is that it eliminates the inter- and intra-observer error associated with human observers. However, it should be noted that at the present time the official protocols are based on human reading and that there are other types of error associated with automated reading. The predicted threshold gold thickness for each detail diameter at each dose level was fitted with a curve as described in the NHSBSP protocol. The confidence limits for the predicted threshold gold thicknesses have been previously determined by a resampling method using a large set of images. The threshold contrasts quoted in the tables of results are derived from the fitted curves, as this has been found to improve the accuracy.5 The expected relationship between threshold contrast and dose is shown in equation 7. Threshold contrast = Dn (7)

D represents the MGD for a 60mm thick standard breast equivalent to the test phantom configuration used for the image quality measurement. is a constant to be fitted. It is assumed that a similar equation applies when using threshold gold thickness instead of contrast. This equation was plotted with the experimental data for each detail size from 0.1 to 1.0mm. The value of n having the best fit to the experimental data was determined. 2.7 Physical measurements of the detector performance
The modulation transfer function (MTF), normalised noise power spectrum (NNPS) and the detective quantum efficiency (DQE) of the CR cassettes were measured using methods as close as possible to that described by the International Electrotechnical Commission (IEC).8 The radiation quality used for the measurements was
adjusted by placing a uniform 2mm thick aluminium filter at the tube housing. The beam quality used was 28kV Mo/Mo. The test device to measure the MTF comprised a 0.8mm thick rectangle (120mm60mm) of stainless steel with polished straight edges. This test device was placed directly on top of the cassette, which was positioned on top of the breast support table. This was positioned to measure the MTF in the laser scan (scan) and plate scan (subscan) directions at three exposure levels. To measure the noise power spectrum, the test device was removed and exposures made for a range of entrance air kerma at the surface of the cassette. The DQE is presented as the average of the scan and subscan directions. 2.8 Image retention
Image retention was measured as described in the NHSBSP protocol and the image retention factor calculated for the three types of plate.

3. RESULTS

3.1 Detector response
Each type of CR plate had the logarithmic response shown in Figure 3ac. The exposure factors and dose levels suggested by the manufacturer resulted in average pixel values of about 18001900 and an entrance air kerma at the cassette surface of about 100Gy. A standard pixel value of 1850 was chosen to determine the reference entrance surface air kerma (ESAK) shown for various beam qualities in Figure 4ac.
y = 405.6Ln(x) 27.4 R2 = 0.998

Average pixel value

Typical values at exposure factors specified by the manufacturer Typical pixel value=1850 ESAK = 102.4 Gy
Entrance surface air kerma at cassette surface (Gy)
Figure 3a Detector response curve for RP-6M at the cassette surface using 28kV Mo/Mo.
y = 403.2Ln(x) 27.4 R2 = 0.998
Typical values at exposure factors specified by the manufacturer Typical pixel value = 1850 ESAK = 102.3 Gy
Figure 3b Detector response curve for RP-7M at the cassette surface using 28kV Mo/Mo.
y = 407.1Ln(x) + 12.6 R2 = 0.999
Typical values at exposure factors specified by the manufacturer Typical pixel value=1850 ESAK = 91.2 Gy

Figure 3c Detector response curve for CP-1M at the cassette surface using 28kV Mo/Mo.
Detector reference air kerma (Gy)

Set tube voltage (kV)

Figure 4a Detector reference air kerma for RP-6M.
Figure 4b Detector reference air kerma for RP-7M.
Figure 4c Detector reference air kerma for CP-1M.
Dose and CNR under AEC control
3.2.1 Dose The mean glandular doses for breasts simulated with PMMA exposed using manual settings designed to achieve the target dose levels proposed by the manufacturer are shown in Table 5 and Figure 5. As the choice of manual mAs settings is limited, the measured doses were slightly different from the target dose levels proposed by the manufacturer. The target and measured doses were generally just below the remedial level in the NHSBSP protocol, which is the same as the maximum acceptable level in the European protocol. The dose for the 6cm thickness was slightly higher than intended and reached the remedial dose level.
Table 5 Mean glandular dose for simulated breasts Equivalent PMMA thickness breast thickness (mm) (mm) kV 29 Measured MGD (mGy) 0.85 1.32 1.72 2.23 2.63 4.50 6.32 Target MGD (mGy) 0.85 1.28 1.70 2.13 2.70 4.28 6.18 NHSBSP remedial level (mGy) >1.0 >1.5 >2.0 >2.5 >3.0 >4.5 >6.5
target Mo Mo Mo Mo Mo Mo Mo
filter Mo Mo Mo Mo Mo Rh Rh

mAs 450

Manual dose setting Remedial dose level MGD (mGy)
Equivalent breast thickness (mm)
Figure 5 MGD for different thicknesses of simulated breasts.
3.2.2 CNR The results of the contrast and CNR measurements are shown in Table 6ac and Figure 6ac. The CNR required to meet the minimum acceptable and achievable image quality standards at the 60 mm breast thickness have been calculated and are shown in Table 4ac and Figure 6ac. The CNR required at each thickness to meet the limiting values for CNR in the European protocol are also shown.
Table 6a Contrast and CNR measurements using RP-6M plates Equivalent breast thickness kV target/ (mm) filter Mo/Mo 26 Mo/Mo 27 Mo/Mo 28 Mo/Mo 28 Mo/Mo 29 Mo/Rh 29 Mo/Rh % contrast CNR at Linearised for minimum CNR at background 0.2mm Measured acceptable achievable pixel value Al CNR IQ lQ 115 21.7 19.9 18.1 16.9 16.3 13.7 12.7 14.15 12.86 10.81 10.17 9.45 8.89 7.86 9.9 9.9 9.9 9.9 9.9 9.9 9.9 14.4 14.4 14.4 14.4 14.4 14.4 14.4 CNR to meet European limiting value 11.3 10.9 10.4 10.2 9.9 9.4 8.9 European limiting values for relative CNR >115 >110 >105 >103 >100 >95 >90
Table 6b Contrast and CNR measurements using RP-7M plates Equivalent breast Linearised thickness kV target/ background (mm) filter mAs pixel value Mo/Mo 26 Mo/Mo 117 % contrast CNR at for minimum CNR at 0.2mm Measured acceptable achievable Al CNR IQ lQ 21.7 19.8 18.1 16.9 16.2 13.6 12.6 13.83 12.34 10.36 9.83 9.09 8.46 7.53 7.6 7.6 7.6 7.6 7.6 7.6 7.6 11.1 11.1 11.1 11.1 11.1 11.1 11.1 CNR to meet European limiting value 8.7 8.4 8.0 7.8 7.6 7.2 6.8 European limiting values for relative CNR >115 >110 >105 >103 >100 >95 >90

27 Mo/Mo Mo/Mo Mo/Mo Mo/Rh 29 Mo/Rh 280 450
Table 6c Contrast and CNR measurements using CP-1M plates Equivalent breast Linearised thickness kV target/ background (mm) filter mAs pixel value Mo/Mo 26 Mo/Mo 116 % CNR at contrast minimum CNR at for Measured acceptable achievable 0.2mm Al CNR IQ lQ 21.5 19.7 17.9 16.7 16.0 13.5 12.4 12.79 11.61 9.69 9.15 8.40 8.33 7.25 5.7 5.7 5.7 5.7 5.7 5.7 5.7 8.3 8.3 8.3 8.3 8.3 8.3 8.3 CNR to meet European limiting value 6.55 6.27 5.98 5.87 5.70 5.41 5.13 European limiting values for relative CNR >115 >110 >105 >103 >100 >95 >90

CNR for 0.2 mm AI

CNR at minimum IQ CNR at achievable IQ CNR to meet European limiting value Measured CNR (manual settings)

PMMA thickness (mm)

Figure 6a Measured CNR compared with the limiting values in the European protocol for the system (error bars indicate 95% confidence limits) for the RP-6M cassette. Using automated IQ measurements to determine the target CNR.
CNR at minimum IQ CNR at achievable IQ CNR to meet European limiting value Measured CNR (manual settings) PMMA thickness (mm)
Figure 6b Measured CNR compared with the limiting values in the European protocol for the system (error bars indicate 95% confidence limits) for the RP-7M cassette. Using automated IQ measurements to determine the target CNR.
Figure 6c Measured CNR compared with the limiting values in the European protocol for the system (error bars indicate 95% confidence limits) for the CP-1M cassette. Using automated IQ measurements to determine the target CNR.

Noise measurements

The variation in noise with dose was analysed by plotting the standard deviation in pixel values against the detector entrance air kerma, as shown in Figure 7ac. If only quantum noise sources were present, the data would form a straight line with an index of 0.5. The presence of structural noise has caused the curve to deviate from a straight line. This is normal for such systems, and the relative contribution from the different noise sources is better understood from the following analysis. The relative noise is plotted against the linearised background pixel value in Figure 8ac. A curve of the form described in equation 5 with an index (n) has been fitted to the measured data in these figures. A value of n of 0.5 would be expected if only quantum noise were present. However, the values of n shown in the graphs (0.33, 0.35 and 0.42) suggest the presence of varying amounts of structural noise. Figure 9ac is an alternative way of presenting the data and shows the relative noise at different average pixel values. The estimated relative contributions of electronic, structural and quantum noise are shown and the quadratic sum of these contributions were fitted to the measured noise using equation 3 and are summarised in Table 7.

Standard deviation in pixel values

y = 0.0899x0.675

Entrance air kerma at surface detector (Gy)
Figure 7a Standard deviation of pixel values versus entrance surface air kerma at the surface of the RP-6M cassette.

y = 0.0880x0.680

Figure 7b Standard deviation of pixel values versus entrance surface air kerma at the surface of the RP-7M cassette.

y = 0.1307x0.595

Figure 7c Standard deviation of pixel values versus entrance surface air kerma at the surface of the CP-1M cassette.

relative noise

y = 0.0821x0.329
Average linearised pixel value
Figure 8a Relative noise at different linearised pixel values for the RP-6M cassette.

y = 0.0911x0.349

Average linearised background pixel value
Figure 8b Relative noise at different linearised pixel values for the RP-7M cassette.

y = 0.134x0.424

BGD pixel value
Figure 8c Relative noise at different linearised pixel values for the CP-1M cassette.
Measured noise fit to noise structural noise quantum noise relative noise electronic noise
detector entrance air kerma (Gy)
Figure 9a Relative noise and noise components at different detector entrance air kerma values for the RP-6M cassette.
Figure 9b Relative noise and noise components at different detector entrance air kerma values for the RP-7M cassette.
Figure 9c Relative noise and noise components at different detector entrance air kerma values for the CP-1M cassette. Table 7 Relative noise components at an entrance air kerma of 100 Gy Plate RP-6M RP-7M CP-1M Total noise (%) 1.68 1.74 1.84 Quantum noise (%) 1.31 1.34 1.65 Structural noise (%) Electronic noise (%) 1.05 1.1 0.8 0.0 0.0 0.0
Image quality measurements
The first exposures of the image quality phantom were made using manual factors to achieve the dose level recommended by the manufacturer for a 5cm thickness of PMMA. This resulted in the selection of 28kV Mo/Mo and 160mAs. Subsequent image quality measurements were made at approximately quarter, half and double this dose by manual selection of the mAs, as shown in Table 8.
Table 8 Images acquired for image quality measurements for each type of CR plate Exposure mode Manual Manual Manual Manual kV target/ filter 28 Mo/Mo 28 Mo/Mo 28 Mo/Mo 28 Mo/Mo Tube loading Mean glandular dose to equivalent Number of CDMAM images (mAs) breasts 60mm thick (mGy) acquired and analysed 0.66 1.32 2.63 5.16
The contrastdetail curves at the four dose levels are shown in Figure 10ac. The threshold gold thicknesses for different diameters and the four different dose levels for each type of plate are shown in Table 9, along with the minimum and achievable threshold values from the NHSBSP protocol (ie same as the European protocol). The data in Table 9 are taken from the fitted curves rather than the raw data. Table 10 shows the predicted threshold gold thickness values determined using automated reading of the CDMAM images. The measured threshold gold thicknesses are plotted against the MGD for an equivalent breast for the 0.1 and 0.25mm detail sizes in Figure 11ac. This shows how the threshold gold thickness reduced as the dose was increased. The fitted curves such as shown in Figure 11ac were used to determine the doses required to meet the minimum acceptable and achievable image quality levels for detail sizes from 0.1 to 1.0mm and are shown in Figure 12ac. Threshold gold thicknesses were determined using human observers and predicted using automatic reading of the CDMAM images.

Figure 10a Contrastdetail curves for the system for four different doses at 28kV Mo/Mo using predicted results from automated reading for the RP-6M cassette. The 2.63 mGy dose corresponds to the initial dose selection proposed by the manufacturer. Error bars indicate 95% confidence limits.
Figure 10b Contrastdetail curves for the system for four different doses at 28kV Mo/Mo for RP-7M plates. The 2.63mGy dose corresponds to the initial dose selection proposed by the manufacturer. Error bars indicate 95% confidence limits.
Figure 10c Contrastdetail curves for the system for four different doses at 28kV Mo/Mo for CP-1M plates. The 2.63mGy dose corresponds to the initial dose selection proposed by the manufacturer. Error bars indicate 95% confidence limits.
Table 9a Average threshold gold thicknesses for different detail diameters for four different doses and human reading for the RP-6M cassette Threshold gold thickness (m)a Diameter (mm) 0.10 0.25 0.50 1.00
Acceptable value 1.680 0.352 0.150 0.091
Achievable value 1.100 0.244 0.103 0.056
MGD = 0.66mGy n/a n/a n/a n/a
MGD = 1.32mGy n/a n/a n/a n/a
MGD = 2.63mGy n/a n/a n/a n/a
MGD = 5.27mGy n/a n/a n/a n/a
These measurements are not yet available.
Table 9b Average threshold gold thicknesses for different detail diameters for four different doses and human reading for the RP-7M cassette Threshold gold thickness (m) Diameter (mm) 0.10 0.25 0.50 1.00 Acceptable value 1.680 0.352 0.150 0.091 Achievable value 1.100 0.244 0.103 0.056 0.4770.048 0.2000.020 0.1090.011 MGD = 0.66mGy MGD = 1.32mGy 1.8200.182 0.3660.037 0.1500.015 0.0820.008 MGD = 2.63mGy 1.3580.136 0.2810.028 0.1120.011 0.0580.006 MGD = 5.27mGy 1.1060.111 0.2120.021 0.0910.009 0.0490.005
Table 9c Average threshold gold thicknesses for different detail diameters for four different doses and human reading for the CP-1M cassette Threshold gold thickness (m) Diameter (mm) 0.10 0.25 0.50 1.00 Acceptable value 1.680 0.352 0.150 0.091 Achievable value 1.100 0.244 0.103 0.056 MGD = 0.66mGy 3.0380.304 0.4800.048 0.1940.019 0.0930.009 MGD = 1.32mGy 2.0570.206 0.3580.036 0.1560.016 0.0790.008 MGD = 2.63mGy 1.2980.13 0.2240.022 0.1010.010 0.0570.006 MGD = 5.27mGy 1.0270.103 0.1900.019 0.0860.009 0.0480.005
Table 10a Average threshold gold thicknesses for different detail diameters for four different doses and automatically predicted data for the RP-6M cassette (data are predicted from automated reading and interpolated using curve fits) Threshold gold thickness (m) Diameter (mm) 0.10 0.25 0.50 1.00 Acceptable value 1.680 0.352 0.150 0.091 Achievable value 1.100 0.244 0.103 0.056 MGD = 0.66mGy 3.2800.496 0.5240.060 0.1940.021 0.1040.013 MGD = 1.32mGy 2.6400.253 0.4010.029 0.1450.010 0.0790.006 MGD = 2.63mGy 1.7600.266 0.2910.033 0.1100.012 0.0610.008 MGD = 5.27mGy 1.4400.138 0.2170.015 0.0970.006 0.0440.003

Table 10b Average threshold gold thicknesses for different detail diameters for four different doses and automatically predicted data for the RP-7M cassette (data are predicted from automated reading and interpolated using curve fits) Threshold gold thickness (m) Diameter (mm) 0.10 0.25 0.50 1.00 Acceptable value 1.680 0.352 0.150 0.091 Achievable value 1.100 0.244 0.103 0.056 MGD = 0.66mGy 2.8970.432 0.4740.039 0.1980.017 0.1130.013 MGD = 1.32mGy 1.9800.199 0.3480.017 0.1430.008 0.0810.006 MGD = 2.63mGy 1.3740.205 0.2600.021 0.1140.010 0.0680.008 MGD = 5.27mGy 1.0110.101 0.1820.009 0.0810.004 0.0460.003
Table 10c Average threshold gold thicknesses for different detail diameters for four different doses and automatically predicted data for the CP-1M cassette (data are predicted from automated reading and interpolated using curve fits.) Threshold gold thickness (m) Diameter (mm) 0.10 0.25 0.50 1.00 Acceptable value 1.680 0.352 0.150 0.091 Achievable value 1.100 0.244 0.103 0.056 MGD = 0.66mGy 2.7720.413 0.3990.032 0.1670.014 0.0980.012 MGD = 1.32mGy 1.8380.184 0.308 0.015 0.132 0.007 0.0720.005 MGD = 2.63mGy 1.0870.162 0.220.018 0.1030.009 0.0590.007 MGD = 5.27mGy 0.880.088 0.1620.008 0.0610.003 0.0310.002
Figure 11a Threshold gold thicknesses at different doses using the RP-6M plates. Error bars indicate 95% confidence limits. The doses indicated are the MGD for a breast equivalent to a 5cm thickness of PMMA.

Figure 11b Threshold gold thicknesses at different doses using the RP-7M plates. Error bars indicate 95% confidence limits. The doses indicated are the MGD for a breast equivalent to a 5cm thickness of PMMA.
Figure 11c Threshold gold thicknesses at different doses using the CP-1M plates. Error bars indicate 95% confidence limits. The doses indicated are the MGD for a breast equivalent to a 5cm thickness of PMMA.

MGD (mGy)

Detail diameter (mm)
Figure 12a The MGD calculated to be necessary to reach the achievable and minimum acceptable image quality levels at different detail sizes using 28kV Mo/Mo for an equivalent breast 60mm thick using the RP-6M cassette. (Derived from automated readings of the CDMAM.)
Figure 12b The MGD calculated to be necessary to reach the achievable and minimum acceptable image quality levels at different detail sizes using 28kV Mo/Mo for an equivalent breast 60mm thick using the RP-7M cassette. (Derived from automated readings of the CDMAM.)
Figure 12c The MGD calculated to be necessary to reach the achievable and minimum acceptable image quality levels at different detail sizes using 28kV Mo/Mo for an equivalent breast 60mm thick using the CP-1M cassette. (Derived from automated readings of the CDMAM.)

Image retention

The image retention factor was very low for all three plates and possibly insignificantly different from zero, as shown in Table 11.

Remedial dose level

Figure 13 Dose to reach minimum acceptable image quality standard for 0.1mm detail. Error bars indicate 95% confidence limits.
Figure 14 Dose to reach achievable image quality standard for 0.1mm detail. Error bars indicate 95% confidence limits.
Figure 15 Dose to reach minimum acceptable image quality standard for 0.25mm detail. Error bars indicate 95% confidence limits.
Figure 16 Dose to reach achievable image quality standard for 0.25mm detail. Error bars indicate 95% confidence limits.

Detector performance

The MTF for the detector system is shown in Figure 17ac for each detector plate and the two orthogonal directions. Figure 18 compares the MTF in the subscan direction for the three detector plates. Figure 19 shows the DQE averaged in the scan and subscan direction with the same dose level (ESAK = 218Gy). Figure 20 shows the DQE for the CP-1M plate at three dose levels.
Figure 17a Modulation transfer function (MTF) for the RP-6M cassette.
Figure 17b Modulation transfer function (MTF) for the RP-7M cassette.
Figure 17c Modulation transfer function (MTF) for the CP-1M cassette.
Figure 18 Modulation transfer function (MTF) measured in the subscan directions using 28kV Mo/Mo for the three detector plates.
Figure 19 Detective quantum efficiency (DQE) (averaged in the scan and subscan directions) using28 kV Mo/Mo and an entrance air kerma of 218 Gy.

110 Gy 218 Gy 442 Gy

Figure 20 Detective quantum efficiency (DQE) averaged in both directions using 28kV Mo/Mo and a range of entrance air kerma for the CP-1M cassette.

4. DISCUSSION

The detector response was, as expected, logarithmic, and it was a straightforward procedure to linearise the pixel values before further analysis. The linearised pixel values were approximately equal to the entrance air kerma at the surface of the cassette when 28kV Mo/Mo was used. Figure 4 showed how the relationship changed at different beam qualities. The noise analysis confirmed that, although quantum noise is the dominant noise source, structural noise was significant for all three types of imaging plate. The relative structural noise ranged from 0.8% to 1.1% for the three types of plate, with the new CP-1M plates having the lowest structural noise. It is one of the characteristics of CR systems that structural noise is not corrected by flat fielding as it would be with a DR system. On the other hand, it was found here that there appeared to be no electronic noise, and this is also a characteristic of CR systems. The dose levels measured were approximately those suggested by the manufacturer and seem to be designed to be just below the limits in the European and NHSBSP protocol. This has the effect of maximising the image quality measurements within acceptable dose limits. It is generally the case that mammography CR systems require higher dose levels to meet the image quality requirements in the NHSBSP and European protocols than do DR systems. The CNR analysis at these dose levels indicated that the currently available RP-6M plates would just fail to meet the requirements in the European protocol (Figure 6a). However, the RP-7M plates passed (Figure 6b). The CP1M plates easily met the minimum standards and reached achievable levels. It is possible that the performance of the CP-1M plate could be further improved by optimising the choice of beam quality and dose. The image quality measurements were conducted at a range of dose levels. The RP-6M plate is within measurement error of the minimum acceptable image quality level at the dose closest to that suggested by the manufacturer (ie 2.63mGy) for the smallest detail size (ie 0.1mm). However, the RP-7M and the CP-1M plates pass at this dose level and could pass at lower dose levels. Thus, the estimated dose required to meet the minimum image quality level for the 0.1mm detail was 2.72mGy, 1.87mGy and 1.47mGy for the RP-6M, RP-7M and CP-1M plates, respectively, using automated reading. Thus, there is a substantial improvement in the performance of the two new plate designs compared with RP-6M. These new plates therefore offer the opportunity either to maintain a high level of image quality or to lower the dose levels while still meeting the minimum image quality level. To put this into context, it appears that the CP-1M plates achieve an image quality/dose balance that is comparable to the Fuji Profect and close to that of film-screen systems. The performance remains inferior to that of modern DR systems, which can either operate at much lower dose levels (by a factor of about 2) or produce better image quality. The measurements of MTF and DQE confirmed the manufacturers claims of improved performance for the two new plates and explain the improved image quality measurements discussed above. The fact that the DQE fell with increasing dose is the usual finding with CR systems and is explained by the presence of structural noise, which becomes more significant at higher doses. The difference between the MTF curves for the scan and subscan directions is also the usual finding for CR systems.

5. CONCLUSIONS

The original RP-6M plate just fails to meet the requirements in the NHSBSP and European protocols. However, both new plates do meet the standards. It would be appropriate to conduct a clinical evaluation of the new plates. From the information available in this study, the CP-1M plate had the best performance. In any clinical evaluation it will be important to verify the robustness of this promising new crystalline technology, eg by inspecting carefully for artefacts.

REFERENCES

1. Workman A, Castellano I, Kulama E, Lawinski CP, Marshall N, Young KC. Commissioning and Routine Testing of Full Field Digital Mammography Systems. NHS Cancer Screening Programmes, 2006 (NHSBSP Equipment Report 0604). 2.Van Engen R, Young KC, Bosmans H, Thijssen M. The European protocol for the quality control of the physical and technical aspects of mammography screening. In: European Guidelines for Breast Cancer Screening, 4th edn. Luxembourg: European Commission, 2006. 3. Alsager A, Young KC, Oduko JM. Impact of heel effect and ROI size on the determination of contrast-to-noise ratio for digital mammography systems. Proceedings of SPIE Medical Imaging, 2008, 6913, 69134I. 4. Young KC, Oduko JM, Bosmans H, Nijs K, Martinez L. Optimal beam quality selection in digital mammography. British Journal of Radiology, 2006, 79, 981990. 5. Young KC, Cook JJH, Oduko JM, Bosmans H. Comparison of software and human observers in reading images of the CDMAM test object to assess digital mammography systems. Proceedings of SPIE Medical Imaging, 2006, 6142, 614206. 6. Young KC, Cook JJH, Oduko JM. Automated and human determination of threshold contrast for digital mammography systems. In: Astley SM, Bradey M, Rose C, Zwiggelaar R, eds. Proceedings of the 8th International Workshop on Digital Mammography. Lecture Notes in Computer Science, 2006, 4046: 266272. 7. Young KC, Alsager A, Oduko JM, Bosmans H, Verbrugge B, Geertse T, Van Engen R. Evaluation of software for reading images of the CDMAM test object to assess digital mammography systems. Proceedings of SPIE Medical Imaging, 2008, 6913, 69131C. 8. IEC 62220-1-2 Determination of the Detective Quantum Efficiency. Detectors Used in Mammography. International Electrotechnical Commission, 2007. 9. Young KC, Oduko JM. Evaluation of Kodak DirectView Mammography Computerised Radiography. NHS Cancer Screening programmes, 2005 (NHSBSP Equipment Report 0504). 10. Young KC, Oduko JM, Woolley L. Technical Evaluation of the Hologic Selenia Full Field Digital Mammography System. NHS Cancer Screening programmes, 2007 (NHSBSP Equipment Report 0701). 11. Young KC, Oduko JM. Technical Evaluation of Kodak DirectView Mammography Computerised Radiography System using EHR-M2 Plates. NHS Cancer Screening programmes, 2007 (NHSBSP Equipment Report 0706). 12. Young KC, Oduko JM. Technical Evaluation of Agfa CR-85 Mammography System. NHS Cancer Screening programmes, 2007 (NHSBSP Equipment Report 0707). 13. Young KC, Oduko JM. Technical Evaluation of the Siemens Novation Full Field Digital Mammography System. NHS Cancer Screening Programmes, 2007 (NHSBSP Equipment Report 0710). 14. Young KC, Oduko JM. Technical Evaluation of the Hologic Selenia Full Field Digital Mammography System with a Tungsten Tube. NHS Cancer Screening Programmes, 2008 (NHSBSP Equipment Report 0801). 15. Young KC, Oduko JM, Gundogdu O, Alsager A. Technical Evaluation of the GE Essential Full Field Digital Mammography System. NHS Cancer Screening Programmes, 2008 (NHSBSP Equipment Report 0803). 16. Oduko JM, Young KC, Alsager A, Gundogdu O. Technical Evaluation of the IMS Giotto Full Field Digital Mammography System with a Tungsten Tube. NHS Cancer Screening Programmes, 2008 (NHSBSP Equipment Report 0804). 17. Young KC, Oduko JM. Gundogdu O. and Alsager A. Technical Evaluation of Sectra MDM-L30 Full Field Digital Mammography System. NHS Cancer Screening Programmes, 2008 (NHSBSP Equipment Report 0805).

Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
NHSBSP Equipment Report 0906 November 2009

Authors

Claire Mercer, Superintendent Radiographer/Breast Screening Programme Manager Dr Anthony Maxwell, Director of Breast Screening both of the Breast Unit, Royal Bolton Hospital NHS Foundation Trust, Minerva Road, Farnworth, Bolton BL4 0JR Mark Worrall, Clinical Scientist Katie Howard, Clinical Scientist both of North Western Medical Physics, The Christie NHS Foundation Trust, Wilmslow Road, Manchester, M20 4BX

Published by

NHS Cancer Screening Programmes Fulwood House Old Fulwood Road Sheffield S10 3TH Tel: 1060 Fax: 1089 Email: info@cancerscreening.nhs.uk Website: www.cancerscreening.nhs.uk The contents of this document may be copied for use by staff working in the public sector but may not be copied for any other purpose without prior permission from NHS Cancer Screening Programmes. The report is available in PDF format on the NHS Cancer Screening Programmes website.
Typeset by Prepress Projects Ltd, Perth (www.prepress-projects.co.uk) Printed by Duffield Printers
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | iii

Contents

Acknowledgements Introduction

v 13 13

2.Objectives of evaluation 3. 4. System description Acceptance testing, commissioning and performance testing
4.1Commissioning 4.2 Six-monthly routine testing 5. Routine quality control
5.1Daily and monthly routine testing 6. Image Quality Assessment
7.Data on screening conducted 7.1Clinical dose audit 7.2Comparison with existing image quality assessment 7.3Conclusions 8.Clinic organisation and throughput 9.Data on assessments conducted 10. Equipment reliability 11.Mammographers comments and observations 11.1Operator manual 11.2Training 11.3 Ease of use 11.4 Image acquisition 11.5 Acquisition workstation 11.6 Image handling 11.7 Reporting workstation

NHSBSP November 2009

iv | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
12. Radiologists/film readers comments 13. Information systems 14.Confidentiality and security issues 15.Conclusions and recommendations References
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | v

Acknowledgements

The authors are grateful to all staff at the Breast Unit, Royal Bolton Hospital NHS Foundation Trust, for their help in the evaluation process. This process could not have been completed without the computed radiography team and experienced film readers and radiologists listed on page 1, or Screening Office Manager Susan Butler. Our thanks are due also to Christine Hopkins for her support and advice throughout the pilot. We are especially grateful for the support of Jenny Diffey and colleagues from North Western Medical Physics, based at The Christie NHS Foundation Trust in Manchester. The contribution of the Konica Minolta team, in association with GE and the Trusts Computer Services Department, was also crucial to the success of the project.
vi | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 1

1. Introduction

This clinical evaluation was undertaken between September 2008 and March 2009 at the Breast Unit, Royal Bolton Hospital NHS Foundation Trust. It was commissioned by Konica Minolta and the NHS Breast Screening Programme (NHSBSP). The NHSBSP Guidance Notes for Equipment Evaluation1 and the guidelines for Routine Quality Control Tests for Full Field Digital Mammography Systems were followed throughout.2 The centre where the evaluation took place is an NHSBSP unit that screens approximately 26000 women per year. It meets both relevant national quality standards for breast screening and the criteria for evaluation centres outlined in the Guidance Notes.1 During the evaluation period the centre was chosen as a pilot site for the extension of the breast screening age. Most of the centres routine screening takes place either on mobile units or in a static screening unit. To achieve the increased capacity needed for the evaluation, more women were screened on the mobile units and at the static site in Bolton. The base site is used primarily for assessment, with only a limited amount of screening. To ensure sufficient screening numbers for the evaluation, however, additional appointments were made there, including some screening outside normal hours. The resulting staffing and other pressures meant that the static site was not an ideal setting for the evaluation; of those available, however, it was agreed to be the most satisfactory. The system was supplied by Konica Minolta on a free loan basis for the evaluation. The company indemnified the equipment for the loan period; it also provided technical support and a laser printer and film for hard copy archiving. The project lead was Claire Mercer, the Breast Screening Programme Manager and an experienced breast screening radiographer. The Director of Breast Screening, Dr Anthony Maxwell, oversaw key clinical decisions. A computed radiography (CR) team was created and its members were the only clinical users of the system under evaluation. The team met monthly for updates on current issues. It comprised Jeanette DouthwaiteTrainee Advanced Practitioner Elizabeth HorneMammographer Advanced Practitioner Elizabeth Hough Jennifer PearceMammographer Alison Tuson Assistant Practitioner The film readers consisted of six experienced radiologists/film readers Advanced Practitioner Christine Crook Dr Janick HarakeConsultant Radiologist Elizabeth Hough Advanced Practitioner Dr Anthony MaxwellConsultant Radiologist/Director of Breast Screening Dr Ann MillsConsultant Radiologist Cheryl Waite Advanced Practitioner North Western Medical Physics (NWMP) collaborated in the commissioning and regular six-monthly testing of the equipment. It also helped with the analysis of clinical evaluation data. The Konica Minolta CR was used with a GE Healthcare mammography unit that was already in place. The collaboration of technical staff from both organisations was vital in setting up the automatic

2 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
exposure control (AEC) to optimise CR performance. The continuing technical support of the Konica Minolta staff was invaluable at every stage of the evaluation process.
2. Objectives of evaluation
The aim of the clinical evaluation was to assess the suitability of the Konica Minolta system for use in NHSBSP mammographic screening. The evaluation used soft copy reporting; the laser printer and hard copy images were used only for archiving purposes. The objectives of the evaluation were to assess clinical image quality by comparing it with the standard local film-screen system to assess the impact on radiation dose of optimising AEC and CR performance to assess the impact of user interfaces on workflow the radiographer workstation and CR plate reader the reporting workstation to determine whether the system could accommodate typical screening appointment slots of six minutes to test the reliability of the system when used for routine NHSBSP screening to evaluate user comments. It was assumed that womens perceptions of the mammography examination would not be affected by the introduction of CR, so these perceptions were not investigated. Radiography questionnaires yielded no comments on the subject.

3. System description

The unit evaluated was the Konica Minolta Regius 190 dual bay CR reader with CS3 console and mammography HQ processing. The mammography processing includes gradation and frequency processing and is supported by Konica Minoltas proprietary Hybrid enhancement processing software. The dimensions of the unit are w580d580h1230 mm and its weight is 170kg. The system uses an RC110M cassette and a rigid CP-1M mammography image plate with two field sizes (1824cm and 2430cm). (See Figure 1.) Its capacity is 90 plates per hour. The cassette cycle lasts approximately 40secs and image review time is 2327secs. The nominal pixel size is 43.75m, with 12-bit (4096) grey levels. The laser imager (which was excluded from the evaluation) uses SDP Dry Laser Film. The workstation hardware comprised Dell Precision WorkStation T5400 two Dual Intel Xeon E5410 processors (each 2.33GHz) 4GB of memory two mirrored 146GB 15K SAS hard drives an NVidia Quadro NVS 290 graphics card for the colour LCD monitor a Dell 2007FP colour monitor a Matrox MDP5MP graphics card for the greyscale monitors two Eizo RadiForce G51 greyscale monitors.
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 3

The software comprised Windows XP Service Pack 3 Insignia InView Standalone QA v5.10.6.
Figure 1 Konica Minolta Regius 190 dual bay CR reader with CS3 console and mammography HQ processing.
4. Acceptance testing, commissioning and performance testing
NWMP were responsible for the commissioning and routine six-monthly testing of the equipment. They also collaborated with the Trust in analysing the clinical data. A Konica Minolta representative was available during the trial period to advise on the operation of the equipment, but took no part in the evaluation process.

Commissioning

NWMP performed commissioning tests on the CR equipment in September 2008. Their report established that new cassettes should be irradiated and erased at least three times before clinical use if 2430cm cassettes are to be used, the AEC device must be adjusted for this larger field size the equipment had met or exceeded the performance standards and protocols set out in NHSBSP Equipment Report 0604 (Version 2).3
Six-monthly routine testing
NWMP performed routine tests on the CR equipment in March 2009, after it had been used for six months in the NHS Breast Cancer Screening Programme. Their report established that there had been no significant changes in performance over the period. Owing to the practical constraints outlined in sections 1 and 8, however, the workload achieved with the equipment was somewhat lower than anticipated.
4 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
5. Routine quality control
5.1 Daily and monthly routine testing
The equipment was not used every day, and tests scheduled to be undertaken daily could be performed only when it was in use. The prescribed daily and monthly tests were undertaken by mammographers. NWMP were provided with the results of AEC consistency tests using 2, 4, 6 and 7cm thicknesses of Perspex to calculate mAs and sensitivity (S-value), which are inversely proportional. The results were compared with standards set out in NHSBSP Equipment Report 0702.2 Using the same target/ filter combination, this gives a remedial level of baseline 10% for the measurement of currenttime and detector dose indicator. The results are given in Figures 26. Figures 25 show that some results were at remedial level for both mAs and S-value. With a 4cm thickness of Perspex, the system selected either the Mo/Mo or Mo/Rh target/filter combination. The results for the two settings have been differentiated in the analysis. Figure 6 shows the variation in S-value after normalising for the variation in mAs. After normalisation, only values for the 4cm block where Mo/Mo was selected remained at the remedial level. As there were few measurement points for this target/filter combination, it was more difficult to define a suitable baseline value for comparisons. In all other cases, the variations in S-value were in the acceptable range. This suggests that the CR system is able to work below the remedial level only if the AEC system operates within smaller tolerances. A region of interest (ROI) is used by the CR system to define the S-value for an image. Differences in the positioning of this ROI influence the S-value, and could underlie the variation in results.

% change from baseline

2cm mAs 6cm mAs 7cm mAs

7/1/09

27/1/09

29/9/08

19/10/08

8/11/08

28/11/08

18/12/08

16/2/09

8/3/09

Figure 2 Percentage change in mAs from baseline for 2, 6 and 7cm of Perspex.
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 5
2cm S-value (linearised) 6cm S-value (linearised) 7cm S-value (linearised)
Figure 3 Percentage change in S-value from baseline for 2, 6 and 7cm of Perspex.

mAs Mo/Mo mAs Mo/Rh

Figure 4 Percentage change in mAs from baseline for 4cm of Perspex.
6 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes
Linearised Svalue (Mo/Mo) Linearised Svalue (Mo/Rh)
Figure 5 Percentage change in linearised S-value from baseline for 4cm of Perspex.

2cm 4cm Mo/Mo

4cm Mo/Rh 6cm

-0.39860 39880

Figure 6 Percentage change in linearised S-value (normalised to baseline value).
6. Image Quality Assessment
Image quality was assessed both objectively, using a test object, and subjectively, using clinical images. For the objective assessment of image quality the CDMAM test object was used. The results were consistent with the NHSBSPs technical evaluation of the system (NHSBSP Equipment Report 0806)4 and (as Figure 7 illustrates) they fell between acceptable and achievable levels. Clinical image quality was tested subjectively by two independent readers. Each assessed 451 images on a five point scale, yielding a total of 902 assessments. The same images were then read and scored by seven advanced practitioners and radiologists.
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 7
predicted data acceptable achievable fit to data Threshold gold thickness (m) 1 Error bars indicate 2 sem

0.01 0.01

0.1 Detail Diameter (mm)
Figure 7 Predicted threshold contrast measurements assessed using the CDMAM test object.
Exposure, contrast and overall image quality were graded from 1 to 5: 1 = inadequate, 2 = poor, 3 = satisfactory, 4 = good, 5 = excellent. Sharpness was graded from 1 to 5: 1 = blurred, 3 = satisfactory, 5 = sharp. Breast composition was defined as fatty, dense or mixed. Three of the images assessed were of patients with implants; four others included unidentified artefacts. There was complete agreement between readers in 49% of cases and agreement to within one point on the scale in a further 48% of cases. As Figure 8 shows, the majority of images were considered good. The sharpness of images was enhanced by the plates needle crystal structure, which reduced scattered radiation.

Number of occurrences

Image quality score 4 5
Exposure Contrast Sharpness Overall
Figure 8 Subjective clinical opinions of screening image quality.
8 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes

Average score

3.0 Dense

Composition

Figure 9 Image quality assessment by breast composition.
In Figure 9 the data have been differentiated according to breast composition. The average sharpness score is slightly higher for fatty breasts than for dense breasts. However the differences are small when compared with the standard deviation in the results.
7. Data on screening conducted

7.1 Clinical dose audit

A dose audit was completed for the first 100 women imaged with the Konica Minolta system. This was based on records of the exposure factors selected by AEC. Four women were excluded from this sample as they had breast implants. The data were analysed using the NHSBSP Breast Dose Calculator and Database (Version 2.0) supplied by the National Coordinating Centre for the Physics of Mammography.5 Figure 10 shows the expected variation in mean glandular dose (MGD) for a range of breast thicknesses. Table 1 shows the mean and range of MGD for both craniocaudal (CC) and mediolateral oblique (MLO) views. The average MGD for MLO view examinations of compressed breasts of 5060mm thickness was 1.98mGy. The standard error of the mean was 0.03 mGy. The mean compressed breast thickness for this sample was 56mm (Table 2). The average MGD for this sample was well below the national reference level, which is 3.5mGy for a compressed breast thickness of 55 mm examined in the MLO view. In 2007 the average MGD for the same GE Senographe DMR+ system using Kodak Min R2000 film was measured at 1.77mGy, with mean compressed breast thickness of 56mm. The increase in MGD is attributable to the adjustment of the AEC for use with a CR system.
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 9
4.3.MGD (mGy) 2.1.0.Compressed breast thickness (mm)

Figure 10 MGD to a range of breast thicknesses.
Figure 10 shows the relationship between MGD and compressed breast thickness. Both CC and MLO views were included in the sample.
Table 1 Average doses for all data in the sample View CC MLO Number of films Mean MGD (mGy) 1.79 2.06 Min max MGD (mGy) 0.78 3.62 0.85 3.82 Mean thickness (mm) 50 55
Table 2 Average doses for 5060mm thick breasts in the MLO view View MLO Number of films 76 Mean MGD (mGy) 1.standard error of the mean 0.06 Mean thickness (mm) 56
Comparison with existing image quality assessment
To assess the Konica Minolta CRs image quality the TOR(MAM) test object was exposed using 28kV, Mo/Mo, and AEC. The AEC registered 129mAs post exposure. This gave a score of 90 on the reporting workstation and 103 on the printed film. By contrast a score of 79mAs post exposure resulted when Kodak Min R2000 film was used on the same x-ray unit. Although the TOR(MAM) had been exposed under the same conditions, the AEC registered 84 mAs post exposure.
10 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes

Conclusions

The average MGD achieved using the GE Senographe DMR+ x-ray set and Konica Minoltas CR equipment (1.98mGy) was slightly higher than that for the same unit using Kodak Min R2000 film (1.77mGy). The increased MGD is attributable to the adjustment of the AEC for use with a CR system. However, it lies comfortably below the National Reference Level of 3.5mGy.6
8. Clinic organisation and throughput
The system was installed in the base screening and assessment unit. The smallest mammography room was used for the CR evaluation, in order to minimise the impact on workflow in the two main symptomatic treatment, screening assessment and biopsy rooms. The monitor for viewing images was sited with the CR reader in the mammography room. The laser printer was needed for archiving purposes as there was no means of archiving to the Picture Archiving and Communications System (PACS). (A mini-PACS has since been installed for use by the breast screening service.) Space being limited, the laser printer was housed in the Screening Programme Managers office. The images were sent to the printer automatically as each examination was completed. The fact that the Konica Minolta CR system was not connected to PACS made some routine activities more laborious. Work-lists could not be downloaded to the workstation, for example. In some cases, mammographers entered each clients identification details during the course of the clinic: in others, details were entered beforehand on a manual work-list. Radiographers were required to record exposure parameters manually in the x-ray rooms. Maintaining a steady throughput of women with consistent appointment times was also a challenge when using the Konica Minolta CR system. However the reasons for this are not related exclusively to the system and its configuration. There was increased throughput at mobile and static sites (away from the base site) to increase capacity in readiness for the extension of the screening age. The base site had not been used every day for screening but kept chiefly for assessment and symptomatic work. Increased screening appointments at the base site demanded more staff, and this had to be achieved without jeopardising the increase in appointments at other sites. Pressures on staffing made six minute clinic appointments impracticable. As a result the number of clients attending a clinic in the evaluation period was somewhat lower than anticipated. Overall throughput for the system under evaluation may have been compromised where mammographers were still engaged at the workstation and not ready to re-enter the x-ray room. However effective teamwork and individual workflow management could help to minimise these delays. (Throughput, and in particular the time taken to digitise the images following the examination, is discussed further in section 11.)

Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 11
To summarise, staff at the site cannot comment authoritatively on the performance of a Konica Minolta CR system in a one-roomed department or mobile screening unit as, for the reasons outlined, it was not possible to evaluate the system at its optimum level. Our experience would nevertheless suggest that typical screening appointment times of six minutes could be supported in a one-roomed static department in which two mammographers were working together.
9. Data on assessments conducted
No breast screening assessments were performed with the Konica Minolta CR system.

10.Equipment reliability

The Konica Minolta CR equipment was generally very reliable. The main challenges arising from the evaluation, and those entailing most system downtime, related to connectivity problems between the acquisition workstation and the printer or reporting workstation. These were not associated with the Konica Minolta CR system however. Uptime cannot accurately be calculated as a proportion of that expected because the equipment was used sporadically. In the event of downtime film-screen mammography could be used; because this alternative was available, repairs might not have been undertaken with the same urgency. The faults directly associated with the CR system were few problems reading the barcodes of cassettes cassettes which occasionally became partly stuck during ejection the difficulties NWMP experienced when loading images on the reporting workstation if a large number of images had been acquired with the same identification. There were also issues after installation with communication errors and links with the laser printer; although internal, these issues were resolved with the help of Konica Minoltas technical staff.
11.Mammographers comments and observations
Mammographers views on the Konica Minolta CR were collected using the relevant form from the NHSBSP equipment evaluation guidelines.1 A summary of their responses to the applicable sections appears below with the project managers observations.
12 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes

Summary of mammographers responses
How good was the operators manual? Satisfactory, although one colleague would have welcomed a printed (as well as the electronic) version. Effective and consistent. The applications specialist was very useful. It was very easy to use, though repetitive strain injury (RSI) risk remains with the handling of cassettes. RSI risk remains with the handling of cassettes. Cassettes were the same size as film cassettes. Generally easy to insert and remove, despite occasional sticking during ejection. (See section 10.) Yes. High. (While some respondents mentioned networking difficulties here, these were caused by internal factors not the CR equipment.) At the start of the trial one user noted that throughput was slower as users were working with an unfamiliar system. At the end of the trial the same user noted that it was possible to maintain the six minute appointment time. Image manipulation enhances film reading. Image quality improved by lack of processing artefacts. Good edge enhancement. Useful to have skin surface and parenchymal patterns visible simultaneously.
How good was the training provided by the supplier? How do you rate the units ease of use?
How do you rate the units help in minimising fatigue? Ease of insertion and removal of cassettes
Were all the expected controls present? What was your level of confidence in the machine? Did the x-ray set performance limit patient throughput?
How would you rate the image quality of films taken on this unit? Additional comments

11.1 Operator manual

The online operator manual was concise and clear. It contained good sequential information and guidance, with recommendations for avoiding potential hazards and system misuse. The Breast Unit received appropriate guidance and support. This included system information to enable us to comply with quality assurance guidelines.

11.2 Training

Training was carried out by an applications specialist from Konica Minolta. This ensured that each mammographer was trained consistently and sufficiently to become competent in all aspects of image acquisition.
Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 13

11.3 Ease of use

Both 1824cm and 2430cm cassettes were used, weighing 400g and 500g respectively. Their loading, scanning, erasing and ejecting did not impede the work flow.

11.4 Image acquisition

This was sufficiently quick to cause no delays.
11.5 Acquisition workstation
The conventional keyboard and mouse and the touch screen workstation were all easy to use, both for generating the work-list and for image acquisition. Combined with automatic image rendering (which helped the mammographer to identify each examination projection) this improved productivity and workflow. It also enabled faster throughput than with film-screen by removing the delay between capturing an image and its appearance on screen. In addition the mammographer was able to create settings for image presentation, which in turn reduced the need for film readers to make adjustments.

11.6 Image handling

A hard copy film was produced for archiving purposes. The time delay in sending images from the acquisition workstation to the reporting workstation was insignificant.
11.7 Reporting workstation
The image viewing application was easy to use. The image processing software expands the image presentation options by allowing parameters such as contrast and brightness to be adjusted independently. Other processing features include the facility to magnify and spot zoom images, and to improve the demonstration of microcalcifications and skin line enhancement in the heterogeneously dense breast. Implant examinations were particularly well demonstrated and enhanced using the image presentation tools.
12.Radiologists/film readers comments
The radiologists/film readers were pleased with the overall image quality. Subjectively, they felt the images compared favourably with the film-screen system in current use. There were occasional technical problems which resulted in the temporary unavailability of screening images. They usually arose from network issues, however, and would be unlikely to occur in a full PACS environment. The reporting workstation provided by Konica Minolta met the needs of the evaluation. Although it had not been set up with specific mammographic display protocols this did not significantly affect the reading and evaluation of the limited number of mammograms in the study.
14 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes

13.Information systems

As noted above, the system was not connected to PACS. Many of the connectivity difficulties experienced would be unlikely to have occurred in a full NBSS/PACS environment.
14.Confidentiality and security issues

Loss of screening images was not a significant risk as images were laser printed for all examinations. Mutual backups of the digital data were provided by image caches on the plate reader and workstation, each cache storing up to 4000 images. This entailed some increased vulnerability to data theft. To enhance data security the system incorporates password protection for all image changes. In addition, the breast unit is secured with digital locks on individual doors and an intruder alarm which is set when the unit is vacated. The digital screening images will be archived to the recently installed screening PACS over the next few months.
15.Conclusions and recommendations
This has been an excellent preparation for the move to digital screening mammography and has given film readers the opportunity to develop their soft copy reporting skills. The Breast Unit is now more aware of where problems are likely to arise and how they can be addressed, with minimal impact on screening activity, when a digital service is introduced. The Konica Minolta CR was used with a GE Healthcare mammography unit already in place; the collaboration of technical staff from both organisations was vital in setting up the automatic exposure control (AEC) to optimise the imaging chain. To evaluate the quality of clinical images, seven experienced readers scored the images on a scale of 1 (inadequate) to 5 (excellent). The majority of images scored 4 (good) in all categories (sharpness, contrast and exposure). The data were separated according to breast composition. Similar image quality was achieved for all breast types. The average sharpness score is slightly higher for fatty breasts than for dense breasts. However the differences are small when compared with the standard deviation in the results. Commissioning tests showed that the equipment was performing to the standards anticipated in the NHSBSP technical evaluation.4 The GE Senographe DMR+ was tested to the NHSBSP standards before the clinical evaluation. Routine tests were completed on both the GE and Konica Minolta equipment after six months; no significant changes in performance were registered. Radiographer quality control tests were completed during the evaluation and satisfactory results were recorded. A dose audit was completed using data from the first 100 women, excluding those with implants. The average MGD for breasts in the range 5060mm was 1.98mGy. This is slightly higher than the

Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes | 15
1.77 mGy measured in 2007 using Kodak Min R2000 film but compares favourably with the national reference level of 3.5mGy. Optimisation has to balance image quality and dose, and a reduction in dose would have adversely affected the image quality. The system under evaluation was very reliable overall. Downtime was caused mainly by connectivity problems between computer systems, although there were occasional difficulties with reading barcodes and with cassettes that became stuck on ejection. There were also issues arising from the organisation of clinics. In particular, staffing pressures made it impossible to schedule six minute appointments, as a result of which the system could not be evaluated in optimum conditions. While this prevents colleagues using it from commenting authoritatively on its capacity, our experience suggests that typical six minute appointments could be supported by this system in a one-roomed static department in which two mammographers were working together. In conclusion, it is the authors view that the Konica Minolta CR 190 performed to acceptable levels in all respects and could be used in the NHSBSP.
16 | Evaluation and Clinical Assessment of the Konica Minolta Regius 190 with CP-1M Cassettes

References

1. Guidance Notes for Equipment Evaluation: Protocol for User Evaluation of Imaging Equipment for Mammographic Screening and Assessment. NHS Cancer Screening Programmes, 2007 (NHSBSP Equipment Report 0703). 2. Routine Quality Control Tests for Full Field Digital Mammography Systems. NHS Cancer Screening Programmes, 2007 (NHSBSP Equipment Report 0702). 3. Commissioning and Routine Testing of Full Field Digital Mammography Systems, Version 2. NHS Cancer Screening Programmes, 2006, (NHSBSP Equipment Report 0604). 4. Technical Evaluation of the Konica Minolta Regius 190 CR Mammography System and Three Types of Image Plate. NHS Cancer Screening Programmes, 2008 (NHSBSP Equipment Report 0806) 5. Patient Dose Software Version 2.0. National Coordinating Centre for the Physics of Mammography. Accessed 3 November 2009 at http://www.nccpm.org/patient_dose_Oz_1.htm. 6. Guidance on the Establishment and Use of Diagnostic Reference Levels (DRLs). Department of Health, 2007. Available at http://www.dh.gov.uk/en/.