In the late 1990s—about 100 years after the discovery of X-rays by Wilhelm Conrad Röntgen—flat detectors were first introduced in medical X-ray diagnostics. Since then this new technology has set out to become the prime standard in digital X-ray imaging. In general projection radiography and mammography flat detectors are increasingly replacing the analog screen-film combinations as well as the digital storage phosphor systems. In cardiac angiography flat detectors with the characteristics of real-time imaging of up to 30 fps (frames per second) or even 60 fps in pediatric cardiac imaging have become so popular that they have almost completely substituted the image intensifier systems within a few years. Changes of this nature are also taking place in general angiography and in fluoroscopy. Hence, flat detectors put us for the first time in a position to finally realize the vision of obtaining a single technology that covers all applications in X-ray diagnostics and interventional techniques.
Any X-ray detector for medical imaging needs to serve the purpose of efficiently absorbing the impinging X-ray flux and converting it into a geometry-conserving digital image signal. The spatial and contrast resolution of the detector should be selected to meet the requirements of the respective application. While the signal per absorbed X-ray should be maximized, the additive electronic noise generated during the numerous conversion and amplification steps is to be kept at a minimum. DQE (detective quantum efficiency) has evolved as the fundamental physical parameter describing the performance characteristics of an X-ray detector as it combines these various requirements. It describes a detector’s ability to convert efficiently the available X-ray radiation at its input into a useful image signal at its output. The objective of any advances in the development of detector technology must be either to reach or to surpass the DQE of previously established detector technologies. A number of design criteria directly influence the main features of a detector. They include active area, pixel size, image acquisition rate, dynamic range, as well as outer dimensions and weight. All of these parameters need to be selected carefully in accordance with the intended application of the detector.
Flat panel detector technology is the only detector technology available today that covers the various requirements of the entire application spectrum, ranging from general radiography to mammography, angiography, and fluoroscopy.
The exciting technology of flat detectors is subject to continuous further development and improvement. Miniaturized electronics allow for even flatter and lighter detectors required for mobile or portable applications. External cooling, still needed in some real-time applications today, will be overcome in the future.
Another area of progress relates to the improvement of the signal-to-noise ratio. This can be obtained through vertical pixel structures. In case of indirectly converting flat detectors, the photodiode and switching transistors are stacked on top of each other instead of being located next to each other as in present and past designs. With such a “maximum fill-factor” approach, the active area of the photodiode is enlarged, allowing to detect nearly all of the generated optical photons and therefore maximizing the signal. The objective of developing new direct conversion materials such as lead iodide, cadmium telluride or mercuric iodide is to improve the absorption characteristics and the overall signal in order to enhance signal-to-noise, resolution and in consequence DQE. Higher bit depths of 16 or 18 bits could improve 3D applications, allowing to probe further into the domain of computed tomography which is the prime standard for low contrast X-ray imaging today.
In the future, pixel arrays based on CMOS (complementary metal oxide semiconductor) technology may be the basis for further progress. Several steps can be envisioned. In the integrating domain, pixel structures including a photodiode and a switch transistor (indirect conversion) or an electrode plus switch transistor (direct conversion) are straight forward designs. Including an amplification stage for each pixel is an improvement which helps to increase the signal-to-noise ratio. One step further could involve counting structures, where each absorbed X-ray quantum is detected individually and registered accordingly. Methods of this kind would allow to generate a quasi-noise-free signal processing chain; only quantum statistics would remain as the sole contributor of noise. Photon counting has already been demonstrated for mammography, even though this approach does not involve a full field detector but a linear detector in a scanning mode. If in addition to counting the quanta, the X-ray’s energy was detected as well, it would be possible to imagine new applications based on energy discriminating methods (“color imaging”). Each new development, be it technical or application-oriented, will be ultimately judged on the basis of its benefit for the patient or the treating physician. Some of the key factors determining the criteria will be workflow, image quality and a careful usage of dose.