User evaluation of detector performance in clinical photon-counting and energy-integrating CT scanners using DICOM images.

Background: Clinical users have a critical need to routinely assess the performance of photon-counting detectors (PCDs) in PCD-CT scanners. Such assessments provide insights into detector characteristics, support protocol optimization, and inform decisions on future scanner acquisitions. Historically, this has been challenging due to limited access to raw detector data, which restricts direct evaluation of PCD performance.

Purpose: To evaluate the zero-frequency detective quantum efficiency ( ${\mathrm{DQE}}_0$ ) and detector deadtime of PCDs from an end-user perspective using reconstructed DICOM images.

Methods: Detector performance was evaluated on two Siemens NAEOTOM Alpha PCD-CT scanners and one Siemens SOMATOM Force energy-integrating detector CT (EID-CT) scanner. Air-only scans were performed in service mode across a range of tube potentials (70-140 kV) and tube currents (4-1200 mA). DICOM images were reconstructed on the scanner using a linear algorithm with a soft-tissue kernel (Br44). The noise power spectrum (NPS) of the images was used to estimate the mean detector output counts. Mean input photon numbers were estimated based on beam quality and exposure measurements. For the PCD-CT systems, tube current-sweep experiments were used to generate image variance-mA curves, from which detector deadtime was estimated using a previously validated parametric model.

Results: The EID and PCD demonstrated comparable ${\mathrm{DQE}}_0$ values (EID: 72%-74%; PCD: 72%-77%). ${\mathrm{DQE}}_0$ showed no significant dependence on tube potential. The estimated PCD deadtime ranged from 5.3 to 7.0 ns. Detector performance was consistent across both PCD-CT systems. Additionally, the ultra-high-resolution (UHR) and standard acquisition modes exhibited equivalent deadtime.

Conclusions: The ${\mathrm{DQE}}_0$ of the PCD-CT and EID-CT detectors are comparable, with performance primarily limited by geometric efficiency rather than sensor absorption efficiency. Under clinically relevant conditions, pileup-induced count losses in the evaluated PCD-CT scanners are minimal and appear to be effectively corrected by the manufacturer.