Photon Counting CT

Level: Introductory Length: 4 hours Format: In-Person Lecture Intended Audience: Scientists, engineers, or managers who wish to learn more about basic strengths and challenges of photon counting detectors for spectral x-ray imaging, how the data is treated and how performance can be quantified. Description: This course explains the principles of photon counting detectors for spectral x-ray imaging. Typical technical implementations are described and fundamental differences to energy integrating systems are pointed out. In particular, the issues of high-rate handling and the effect of detector cross talk on energy resolution are described. Requirements on electronics for spectral imaging in computed tomography is also discussed. A second objective of the course is to describe how energy sensitive counting detectors make use of the energy sampling of the linear attenuation coefficients of the background and target materials for any given imaging task; methods like material basis decomposition and optimal energy weighting will be explained. The second objective highlights the interesting fact that while the spatial-frequency descriptor of signal-to-noise-ratio transfer (DQE) of a system gives a complete characterization of performance for energy integrating (and pure photon counting) systems, it fails to characterize multibin systems since a complete description of the transfer characteristics requires specification of how the information of each energy bin is handled. The latter is in turn dependent on the imaging case at hand which shows that there is no such thing as an imaging case independent system DQE for photon counting multibin systems. We also suggest how this issue could be resolved. Learning Outcomes: This course will enable you to: - describe the fundamental operating principles of photon counting detectors for spectral x-ray imaging - distinguish between the proposed detector materials in terms of their main physical limitations/challenges to high-rate energy resolved photon counting - list essential requirements on read-out electronics and predict effect on image quality if not fulfilled - explain the physical origin of pile-up and separate between the effects of decreased energy resolution and loss of counts - explain the physical origins of cross-talk and how it degrades performance, both in terms of resolution and noise - compute optimal weights for the energy bins - illustrate how poor choice of weights results in inferior image quality - perform material basis decomposition and explain why noise in decomposed images is a poor figure-of-merit - distinguish between system DQE and task dependent DQE and suggest solutions to allow comparison at system level between multibin energy resolved systems and other solutions Instructor(s): Mats E. Danielsson has been developing photon counting x-ray detectors for medical imaging for 15 years and his research has resulted in detector systems in worldwide clinical use. He received his Ph.D. in experimental physics in 1996 based on work at CERN, Geneva and later did his postdoc at Lawrence Berkeley National Laboratory. In 2006 he was appointed Professor at KTH Royal Institute of Technology in Stockholm, Sweden, where he heads the physics of medical imaging research group. Dr. Danielsson is a lifetime member of SPIE. Mats U. Persson received his MSc in Engineering Physics in 2011 and his PhD in Physics in 2016, both from KTH Royal Institute of Technology in Stockholm. His PhD work was centered on photon-counting spectral CT imaging with a photon-counting silicon strip detector. After working as a postdoc at Stanford University and as a visiting postdoc at General Electric Research Center, he returned to KTH in 2020 where he is now an Assistant Professor of Physics. His research interests are focused on image reconstruction and mathematical performance modeling for photon-counting spectral CT. Event: SPIE Medical Imaging 2025 Course Held: 16 February 2025