The X-ray spectrum from a conventional tungsten target, glass encapsulated, aluminium filtered X-ray tube is not optimal for mammography. The best subject contrast for between normal and malignant tissue is considered to be produced at a photon energy of around 20 keV, which is much lower than that used in normal radiography. Increasing photon energy will reduce contrast and reducing photon energy will lead to inadequate penetration of the breast and a large increase in patient dose, so the X-ray spectrum is critical. A range of mammographic spectra are used for digital mammography. The X-ray tube target may well be switchable (depending on the design) between molybdenum and rhodium, or rhodium and tungsten and the tube has a low attenuation beryllium output window. The beam is then filtered with either molybdenum, rhodium, silver or aluminium filters. The X-ray tube is operated at a voltage in the range 25–35 kV.

Figure 16.2 shows the spectrum of a rhodium target, rhodium filtered beam at a tube voltage of 30 kV. Rhodium has characteristic X-ray peaks at 20.2 and 22.7 keV, which contribute strongly to the limited range spectrum. Rhodium is again used as the filter because, due to the K-edge absorption, it strongly attenuates energies just above its own K-characteristic peaks as well as attenuating lower energies. The end result is a spectrum with most photons lying in a narrow band of energies.

Although the most common spectrum, this is not necessarily optimised for all breast thicknesses. For larger breasts, a more penetrating beam is optimal to avoid very long exposure times which bring the possibility of movement blur, tube overloading and high radiation doses. Figure 16.3 shows the spectrum of a tungsten target, aluminium filtered beam again at a tube voltage of 30 kV, with the X-ray tube again having a beryllium output window. Although the tungsten anode, aluminium filter combination was not used with film-screen mammography, it is well suited to the response of modern digital mammography receptors and is becoming a more common choice.

The shape of the spectrum is quite different from Fig. 16.2 and its peak is now at a higher energy, even though the tube voltage is the same. The tungsten target has no K-characteristic X-ray peaks in this energy range, and the aluminium filter, which similarly does not have a K-absorption edge in this energy range, does not preferentially attenuate the higher energy end of the spectrum.

In mammography the breast is compressed using a rigid transparent plastic compression plate which can be motor driven. The use of compression force reduces the thickness of the breast and holds it in place which gives a number of advantages:

Compression in mammography is one of the few occasions in radiography where a technical advantage is gained without detriment to other aspects of the image; although there is a disadvantage in client discomfort. Modern mammography units can employ a system to measure the increasing amount of force resulting from a given small increase in compression to stop the motorised movement at a given compression. Many units use a motor driven assembly with more or less compression being applied by the practitioner; the practitioner has direct control over the amount of compression applied. It is suggested that the compression paddle must be controlled by hand during the final compression [8].The compression force maximum limit set on mammography systems is 200 Newtons. A range of compression paddles are normally supplied with a digital mammography unit to cover different types of projection. Some typical types are:

Digital image capture was first introduced into mammography as ‘small-field digital mammography’ for needle and core biopsy guidance using detectors typically approximately 15 cm in size. Full-field digital mammography, with detector sizes up to the equivalent of the 24 × 30 cm was developed later.

The imaging advantages of digital mammography include a wide dynamic range and the separation of the image capture and image display functions, so that the image display can be varied to optimally show the full range of recorded X-ray intensities. This provides good visualisation of the skin line and nipple and has advantages when imaging dense breasts and younger women. There are a range of competing image capture technologies for digital mammography, and the choice of technology depends to some extent on the imaging task, be it screening or symptomatic, and the size and format of the hospital or clinic environment in which it will be operated. The range of technologies described below fit the state of the market at the time of writing, but this is an area where progress is still rapid.

A digital image is not a continuous distribution of bright and dark, but is composed of a finite number of points (or ‘pixels’), where each pixel has a value of brightness dictated by a stored numerical value. Digital images have the advantage that they can be enhanced and manipulated by computer to extract the maximum amount of diagnostic information. Digital images can be stored, transferred, copied without detriment and retrieved in a very efficient manner using computer mass data storage techniques. They have the disadvantage of a limit to spatial resolution caused by the finite pixel size. Digital mammography receptors usually have a linear response between pixel value and the radiation dose incident on the pixel over a very wide dynamic range, typically a factor of some 10,000:1. The choice of what patient dose is required for digital mammography is therefore driven by the signal-to-noise ratio required for a diagnostic image rather than a specific radiation dose to the receptor.

In direct conversion detectors the X-ray interaction is converted directly to an electrical signal using an amorphous selenium (a-Se) layer, behind which lies an amorphous silicon micro-circuit layer which in turn is supported by a rigid substrate (Fig. 16.7). Selenium is a photoconductor, so is an electrical insulator in the dark, and a conductor when exposed to light or X-rays. The amorphous selenium is employed as a mammographic image receptor in the form of a thin layer (0.5 mm) with a voltage applied between a large area electrode across the front surface, and an array of charge collection electrodes, one per pixel, on the back surface. These are linked to capacitors to accumulate the charge released during the exposure. These are linked in turn to thin-film-transistor switches to provide a line-by-line read out arrangement in which the charge stored for an individual pixel is passed pixel-by-pixel along the line until it can be measured by electronics external to the imaging sensor. Incoming X-ray photons interact photo-electrically in the a-Se layer producing electrons and ‘holes’ (the vacancy where an electron should be). Because of the high voltage gradient across the thin a-Se layer, the electrons move towards the positive surface electrode and the holes towards the negative charge collection electrodes. The electrons and holes do not move sideways as they have to follow the direction of the electric field gradient, so image blurring from this source is minimal and the spatial resolution of the detector is good. At the end of the exposure, the charge signals (proportional to the radiation detected) from each pixel are read out via the thin-film-transistor switches and data lines. The charge signals are converted to digital values via charge amplifiers and with a digital-to-analogue converter and sent to the computer for assembly into an image.