Digital mammography: latest advances

Magnus Åslund
PhD
Department of Physics
Royal Institute of Technology
AlbaNova Stockholm­
Sweden

A decade ago, full-field digital mammography was a promising new technology. Recently it was shown that, for breast cancer screening, digital mammography has significantly greater diagnostic accuracy than film for women with dense breasts, women under 50 years, and pre- and peri‑menopausal women. For the entire population studied, digital was equivalent to film.(1) The likely reason for the improved accuracy in women with dense breasts is the separate acquisition and display. It allows image processing to visualise small contrasts (eg, between a cancer and adjacent dense tissue). For the majority of the digital systems studied, the improved performance was achieved at a lower dose than film, probably due to the superior detector performance of the digital systems.(2)

Digital projection mammography
The chronology of technologies currently available in Europe starts with indirect computed radiography (CR), followed by an indirect flat panel using CsI (caesium iodide) as scintillator, direct a-Se (amorphous selenium) flat panels and a direct photon-counting system using Si (silicon).(3) These systems use a grid, with the exception of the photon-counting system, which uses a scanning ­multislit geometry. The general trend in this development is to detect more photons and have fewer noise sources in the detection chain. In addition, each detector technology improves with time by engineering (eg, dual-side read CR and an indirect CsI flat panel with lower detector noise).(4,5) Comparisons against film have shown matched image quality at 30–50% of the dose for the indirect CsI system.3 Efficient scatter rejection allows for a further 30–50% dose reduction at maintained image quality, which has been ­demonstrated for the scanning multislit system.(6,7)

Minimising radiation risk is important in general, as manifested by the ALARA (“as low as reasonably achievable”) principle. In ­mammography, radiation risk is relatively low but still a factor in the benefit/risk ratio of screening.(8) In phantom studies, increasing the dose (through the mAs) results in improved ­image quality.(7) However, there is to the author’s knowledge no evidence that an increased phantom image quality, compared to the current levels that are used in mammography, would lead to improved diagnostic accuracy when screening for breast cancer. Further improvements in system performance should then be used to lower the dose at maintained clinical performance.

Imaging techniques in digital mammography
There is room for improving diagnostic accuracy; in projection mammography, 10–20% of breast ­cancers are not detected, and only 5–40% of detected lesions recommended for biopsy are malignant.(9) The ­anatomical background noise created by superimposed tissue has been proposed as one limiting factor.

Tomosynthesis
Three-dimensional imaging could reduce the superimposed anatomical noise intrinsic to projection ­mammography. Tomosynthesis reconstructs a volume from a series of projection images taken at different­ angles.(10) Prototypes for mammography have been developed with the available ­detector technologies (see Figure 1). The exception is CR, which lacks the required readout rate. These prototypes sample 11–21 projection images spread in a 12–60° angle by moving the X-ray tube in an arc above the ­detector. The reconstructed volume has high in-plane resolution but relatively poor depth resolution due to the limited angle. Acquisition times are 8–15 seconds, and the dose is comparable to that used for one to two conventional images. Among the challenges are the dose per projection versus detector noise, the high readout rate versus lag effects, long acquisition times versus motion blurring, and the incomplete data causing artifacts. Initial clinical studies have shown improved image quality in the tomosynthesis slices and increased specificity.(11) In a study using phantoms with simulated anatomical noise, the accuracy of projection mammograms could be matched at 25% of its dose.12 Clinical evaluations are underway.(12,13)

[[HHE07_fig1_R22]]
 
Subtraction mammography
Neoangiogenesis creating leaky vessels accompanies tumour growth above a few ­millimetres. Contrast-enhanced techniques observe the ­kinetics, intensity and morphology of a contrast agent administered intravenously, which, due to the neoangiogenesis, concentrates and enhances tumours. Clinical ­feasibility studies have been performed in combination with temporal or energy subtraction in order to suppress the anatomical background.(14,15) In the former case, a precontrast image is subtracted from several post-­contrast images. The dose is comparable to that used for one conventional image. In the dual-energy study, two postcontrast images at different X-ray energies are subtracted (see Figure 2). The dose was 0.7 mGy higher than that used for the conventional image. Among the challenges is sensitivity to reproducibility of, for example, breast position between exposures (eg, the temporal may suffer from the injection time). A phantom study of the energy subtraction method has also been performed using an energy-sensitive photon-counting detector generating the high- and low-energy image from a single exposure.(16)

[[HHE07_fig2_R22]]

Clinical feasibility studies show that contrast-enhanced subtraction could demonstrate cancers not visible in projection mammography and potentially differentiate between benign and malignant ­lesions.(15,17) Research is now initiated on contrast-enhanced subtraction methods in combination with tomosynthesis.(18)

Digital mammography can provide equal or greater diagnostic accuracy at a lower dose compared with film in screening for breast cancer. However, increasing diagnostic accuracy further by means of dose seems difficult. Tomosynthesis and contrast-enhanced digital subtraction are techniques using the platform of digital ­mammography that could potentially improve diagnostic accuracy.

References

1. N Engl J Med 2005;353:1773-83.
2. Med Phys 2006;33:719-36.
3. Eur Radiol 2006;16:38-44.
4. Med Phys 2003;30:1843-54.
5. Proc SPIE 2005;5745:1078-86.
6. Med Phys 2006;33:933-40.
7. Proc IWDM 2006;4046:441-6.
8. Radiat Prot Dosimetry 2005;117:318-20.
9. Radiographics 2004;24:1747-60.
10. Phys Med Biol 2003;48:65-106.
11. Proc IWDM 2006;4046:152-9.
12. Proc IWDM 2006;4046:160-6.
13. Proc IWDM 2006;4046:137-43.
14. Radiology 2003;228:842-50.
15. Radiology 2003;229:261-8.
16. Eur J Radiol 2006;60:275-8.
17. Invest Radiol 2005;40:397-404.
18. Proc IWDM 2006;4046:183-9.