Photon-counting Detectors in Clinical Computed Tomography

Stefan Ulzheimer, Steffen Kappler|2017-01-13

The detector system is arguably the key component in a computed tomography scanner. Its task is to convert the incoming X-rays into an electrical signal that can then be fed into the image reconstruction chain. Decisive properties are high quantum efficiency (X-ray stopping power), large dynamic range, and fast signal decay (low afterglow) combined with excellent signal stability.


CT detectors have undergone many development steps from the very early gas detectors – sometimes radical, sometimes merely incremental. Based on its proprietary scintillator material UFCTM, Siemens continuously sets new standards in CT detector design.[1] The latest innovation is the fully integrated StellarInfinity detector that Siemens uses in today’s high-end CT systems (Fig. 1), first introduced with SOMATOM Force. Due to its unprecedented level of electronic integration, the StellarInfinity detector delivers exceptionally low levels of electronic noise and high signal dynamics. This makes it the optimal tool to exploit current and future iterative reconstruction possibilities.

Today’s solid-state detectors use a two-step process to convert the incoming X-ray intensity into an electronic signal. First, the X-rays are converted into visible light in a scintillator layer. Below, the figure shows a photodiode array that converts the emitted light into an electric current that is digitized in dedicated ASICs. The scintillator layer is made of a ceramic material that is mechanically structured into pixels. Septa of finite width separate the individual pixels in order to suppress optical cross-talk (Fig. 2A).

If such detectors, however, do become available, they will offer tremendous advantages. It would not only be possible to register or count each individual photon but also to measure the energy from each individual photon. Without any other technical prerequisites, multi-energy CT could become possible. In principle, as a next step following on from dual energy CT, an arbitrary number of energies is achievable. Still, in order to limit the amount of data delivered by the detector, it is more practical to put the registered photons within certain energy ranges or “energy bins” and only read the data for these bins. Today’s dual energy CT is sufficient for the materials present in the human body and for iodine contrast. More “bins” would not add more information here. Should new contrast materials become available in the future, however, such as gold nanoparticles with additional absorption peaks in the used X-ray energy range,[2] three, four, or even more, bins might be beneficial (k-edge imaging).

Since only signal pulses above a certain threshold are registered in photon-counting CT scanning, such detectors completely eliminate electronic noise that is produced in every electronic component until a signal is digitized. The StellarInfinity detector set a new standard with exceptionally low electronic noise due to its complete integration of photodiodes and analog-digital convertors in one single chip. There are clear advantages in clinical scenarios of using extremely low doses [3] or when having to deal with low signals such as in obese patients. Photon-counting detectors will take this to a new level. In addition, photon-counting detectors are able to weigh low-energy photons higher than high-energy photons. This allows even further optimization of the contrast-to-noise ratio (CNR) in contrast-enhanced applications. Initial studies show that an additional dose reduction of 32% will be possible.[4]

In today’s scintillators, the pixels have to be mechanically structured and therefore cannot be infinitely small. With decreasing pixel size, the geometrical dose efficiency drops because the relative area of the septa between the pixels increases. In photon-counting detectors, a mechanical separation of the pixels is not necessary. Pixels can be structured in a photolithographic process as already established for silicon chips. In principle, pixels can be made very small. Of course, an optimized value must also be found here that delivers practical additional clinical value. Recent publications in the field of computed tomography indicate that the pixel pitch in photon-counting detectors for clinical scanners might be reduced by a factor of three to four relative to today’s solid-state detectors.

Siemens is the first medical equipment company to be able to demonstrate the potential of high-flux capable photon-counting technology in two near-clinical prototype installations.[5,6] The prototypes are based on SOMATOM Definition Flash hardware in which the second detector system was replaced with a photon-counting detector. For the first time, clinical equivalence comparable to existing technology was demonstrated in human subjects. In the next five years, Siemens will collaborate with its network of partners to optimize the technology and to identify new clinical potentials.


Stefan Ulzheimer, PhD, Siemens Healthineers, Germany
Steffen Kappler, PhD, Siemens Healthineers, Germany

1