Speaker
Description
Introduction
Computed Tomography (CT) is an integral part of contemporary medicine. Conventional CT scanners are equipped with Energy Integrating Detectors (EID), which are based on indirect conversion of the incoming photons into electric signal for the image generation. EID-based CT shows limitations regarding tissue contrast and spatial resolution. A turning point of the actual CT technology is reached by Photon-Counting CT (PCCT), which embodies Photon-Counting detectors (PCDs) in substitution of EIDs. PCDs are characterized by a semiconductor layer, usually cadmium telluride (CdTe) or cadmium zinc telluride (CZT), where x-ray photons are absorbed, generating electron-hole pairs that are collected by pixelated anodes. The resulting signal is proportional to the absorbed photon energy and a reading logic based on a calibrated threshold permits to distinguish and count detected photons. The final pixel value reflect the number of detected photons and the use of multiple thresholds permits also to recover information related to the energy of such photons. By means of dedicated software, this allows the generation of virtual monocromatic images, at all the energies of the incident X-ray spectrum.
State-of-the-art clinical applications of PCCT regard cardiovascular exams, thoracic and lung imaging, and neuroimaging [1].
In this work, we characterized the first clinical PCCT in terms of spatial resolution performances. The scanner is the Naeotom Alpha® (Siemens Healthineers equipped with CdTe PCDs
Materials and methods
Image quality was assessed by investigating spatial resolution properties of the PCCT scanner. We measured the high-contrast pre-sampled Modulation Transfer Function (MTF) that quantifies the system response to an impulsive input in the transformed Fourier domain, bearing information regarding the signal transfer across the spatial domain [2]. We acquired a tungsten wire (purity of 99.95%) with a diameter of 12.5 μm. The wire was located both at the isocenter of the CT gantry and at lateral positions (4 cm horizontally and 5 cm vertically) in the axial plane (Fig. 1). Images were acquired using Inner Ear clinical protocol, at 140 kV and 80 mAs (these parameters were chosen based on routine clinical practice). Images were acquired with helical scans with pitch of 0.85.To obtain the pre-sampled MTF, the wire was positioned slightly angled in the coronal plane (approximatively 2-3°). Acquisitions were made both at conventional (144x0.4mm) and at ultra-high resolution (120x0.2mm) mode, resulting in different reconstructed slice thicknesses. 0.4 mm - slice images are 640x640 matrix with a pixel size of 0.3125 mm, while 0.2 mm - slice images are 1024x1024 matrix with pixel size of 0.1953 mm. We reconstructed the final images using multiple kernels, chosen among those available for standard and high-resolution inner ear exams (Hr72, Hr80, Hr98, and Hv89). In detail, Hr72, Hr80, and Hr98 are kernels used for head exams, while Hv89 is used for vascular protocols. Increasing the kernel index means improving spatial resolution but also increasing image noise [3].
Results
Spatial resolution depends on the reconstruction kernel and the position of the test object within the gantry in the axial plane. Spatial resolution is quantified by means of the spatial frequencies f50% and f10%, corresponding to 50% and 10% values of MTF maxima. Figg. 2,3,4 show MTF curves for the three positions of the tungsten wire. Results for f50% and f10% are also collected in Tab. 1.
Standard reconstruction kernels with higher index (Hr80 and Hr98) show higher MTF parameters, as expected. MTF curves and parameters calculated for the displacement in x and y direction differ from those found at isocenter. This could be attributed to the x-y length of the focal spot of the system as well as to the reduced sampling for peripheral portion of the FOV.
Conclusions
This work aimed to provide a characterization of the spatial resolution of a PCCT scanner used in clinical practice. First results show that iterative reconstruction kernels impact axial-plane resolution, imparting non-isotropic 3D resolution characteristics.
In addition, MTF curves vary across the radial (x-y) directions in the axial plane. Further investigation will involve an improved experimental setup and more acquisition of the tungsten wire at added locations in the scanner gantry, also to characterize longitudinal and azimuthal spatial resolution of the CT system.
References
[1] M. Tortora, L. Gemini, I. D’Iglio, L. Ugga, G. Spadarella, and R. Cuocolo, ‘Spectral Photon-Counting Computed Tomography: A Review on Technical Principles and Clinical Applications’, J. Imaging, vol. 8, no. 4, p. 112, Apr. 2022, doi: 10.3390/jimaging8040112.
[2] H. Fujita et al., "A simple method for determining the modulation transfer function in digital radiography," IEEE Trans. Med. Imaging, vol. 11, no. 1, pp. 34-39, 1992, doi: 10.1109/42.126908.
[3] H. Onishi et al., "Photon-counting CT: technical features and clinical impact on abdominal imaging," Abdom. Radiol., vol. 49, no. 12, pp. 4383-4399, 2024, doi: 10.1007/s00261-024-04414-5.
| Workshop topics | Applications |
|---|
