Click here to

Session: Performance Assessment and Clinical Implementation of Photon-Counting CT [Return to Session]

Incorporating a Photon Counting CT System Into Routine Clinical Use

S Bache*, J Solomon, E Samei, Duke University, Durham, NC


TH-D-201-5 (Thursday, 7/14/2022) 11:00 AM - 12:00 PM [Eastern Time (GMT-4)]

Room 201

Purpose: To determine a methodology for incorporating a photon counting CT system into routine clinical practice by comparing image quality and radiation dose metrics to a traditional CT system

Methods: An ACR-CT phantom was scanned on a conventional energy-integrating (EID CT) and a Siemens photon-counting (PCCT) CT system at three dose levels, and reconstructed with most available convolution kernels and iterative reconstruction strengths. Image sharpness was characterized by the 50% cutoff frequency (f50) of a clinically relevant TTF and noise texture was characterized by the average frequency (fav) component of the noise power spectrum in a uniform water section of the phantom. Image perception was matched by seeking the most closely positioned kernels from each scanner type on a plot of f50 versus fav. In addition, CTDIvol necessary to obtain a noise value of 5 HU in uniform water images (dose5) was computed using a log-linear fit of noise versus acquisition CTDIvol.

Results: For clinically routine kernels, identically named PCCT kernels were found to be 5-10% less sharp than EID CT counterparts. Noise texture comparison followed from sharpness findings, with EID CT reconstructions having increased fav values compared to the PCCT counterpart. The PCCT implementation of iterative reconstruction (QIR) showed a better preservation of noise texture as iterative reconstruction strength was varied from “off” to maximum strength compared to EID CT. Over all kernels, dose5 values were 5-10% lower on PCCT reconstructions compared to EID CT.

Conclusion: Image perception metrics from PCCT and EID CT scanners were used to inform protocol design while incorporating a PCCT system into routine clinical use. These metrics showed that routine EID CT kernels should be increased by one sharpness step, maximum or near-maximum iterative reconstruction strengths could be used while preserving noise texture, and doses 5-10% lower than EID CT could be used.


Computed Radiography, Photon Detectors


IM- CT: Quality Control and Image Quality Assessment

Contact Email