Title page
Contents
About Public Health England 2
About PHE Screening 2
Executive summary 4
1. Introduction 5
1.1. Testing procedures and performance standards for digital mammography 5
1.2. Objectives 5
2. Methods 6
2.1. System tested 6
2.2. Dose and contrast-to-noise ratio using AEC 8
2.3. Image quality measurements 10
2.4. Geometric distortion and reconstruction artefacts 10
2.5. Alignment 11
2.6. Image uniformity and repeatability 12
2.7. Detector response 12
2.8. Timings 12
2.9. Modulation transfer function 13
2.10. Local dense area 13
3. Results 15
3.1. Dose and contrast-to-noise ratio using AEC 15
3.2. Image quality measurements 17
3.3. Geometric distortion and resolution between focal planes 19
3.4. Alignment 21
3.5. Image uniformity and repeatability 21
3.6. Detector response 22
3.7. Timings 22
3.8. Modulation transfer function 22
3.9. Local dense area 24
4. Discussion 26
4.1. Dose and contrast-to-noise ratio 26
4.2. Image quality 26
4.3. Geometric distortion and reconstruction artefacts 26
4.4. Alignment 27
4.5. Image uniformity and repeatability 27
4.6. Modulation transfer function 27
4.7. Local dense area 27
5. Conclusions 28
References 29
Table 1. System description 6
Table 2. Radiographic factors for pre-pulse exposures, selected according to compressed breast thickness (CBT) 7
Table 3. Image file sizes for 60mm CBT, 24cm x 29cm field size 7
Table 4. HVL and tube output measurement in tomosynthesis mode 15
Table 5. Dose for tomosynthesis images acquired using AEC 16
Table 6. CNR for tomosynthesis images acquired using AEC 17
Table 7. Threshold gold thickness for reconstructed focal plane 54 and slab 5 of the image of the CDMAM phantom (automatically predicted data) 19
Table 8. z-FWHM measurements of 1mm diameter aluminium balls 21
Table 9. Scan and reconstruction timings 22
Table 10. MTF for central projections in the directions parallel (u) and perpendicular (v) to the tube axis 23
Table 11. MTF50 for central projection 23
Table 12. AEC performance for local dense areas, measured on the midline and 50mm from the CWE 24
Figure 1. The GE Senographe Pristina digital breast tomosynthesis system (image courtesy of GE Healthcare) 8
Figure 2. The position of 5mm x 5mm ROIs for assessment of CNR (The CWE is to the left) 10
Figure 3. Set-up to measure AEC performance for local dense areas 14
Figure 4. MGD for tomosynthesis exposures acquired using AEC. Error bars indicate 95% confidence limits 15
Figure 5. CNR for tomosynthesis images acquired using AEC. Error bars indicate 95% confidence limits 16
Figure 6. Variation of projection CNR with angle for images of 45mm PMMA. Error bars indicate 95% confidence limits 17
Figure 7. Threshold gold thickness for plane 54, at 3 dose levels. Error bars indicate 95% confidence limits 18
Figure 8. Threshold gold thickness for slab 5, at 3 dose levels. Error bars indicate 95% confidence limits 18
Figure 9. Appearance of 1mm aluminium balls in reconstructed focal planes at 1mm intervals, from 4mm below to 3mm above the plane of best focus 20
Figure 10. Appearance of 1mm aluminium balls in reconstructed slabs at 5mm intervals, from 5mm below to 10mm above the slab of best focus 20
Figure 11. Extracts from planes showing 1mm aluminium ball in (i) single focal plane, (ii) Maximum Image Projection (MIP) through all focal planes, and through re-sliced vertical planes in the directions (iii) parallel and... 21
Figure 12. Detector response in tomosynthesis mode 22
Figure 13. MTF for central projections 23
Figure 14. AEC performance in projection images for local dense areas 25