Clinical Photon-Counting CT in Neuroradiology: A Real-World Paired Comparison with Conventional Energy-Integrating CT

Photon-counting CT offers higher spatial resolution, improved contrast-to-noise efficiency, and spectral imaging, but its performance against heterogeneous EID-CT systems in routine use remains incompletely characterized. We compared image quality, pathology conspicuity, and radiation dose between clinical PCCT and EID-CT in paired routine neuroradiological examinations, accounting for differences in scanner generation, protocol, and reconstruction. We hypothesized higher image-quality and pathology-conspicuity scores and dose efficiency, while treating the comparison as pragmatic rather than detector-isolating.

This retrospective study included consecutive adults undergoing routine clinical PCCT from December 2023 through February 2024 who had a corresponding same-region EID-CT within three months; when several EID-CT scans were eligible, the closest in time was selected. Examinations were grouped by protocol: non-contrast brain CT was the primary cohort; contrast-enhanced brain CT and arterial and venous CTA were exploratory. Four blinded readers assessed image quality and pathology conspicuity, with paired scans assigned to separate sessions at least four weeks apart. CNR, dose-normalized CNR (CNRD), and radiation dose were compared within subjects using the Wilcoxon signed-rank test.

The 194 patients (mean age, 65±15 years; 100 males) contributed to 270 pairs: 225 non-contrast brain CTs, 14 contrast-enhanced brain CTs, 26 arterial CTAs, and 5 venous CTAs. In the primary cohort, PCCT was associated with higher overall image-quality scores (median 4 versus 3; p<.001), higher time-independent pathology conspicuity in 224/225 pairs (median 4 versus 3; p<.001), 102%-152% higher GM/WM CNR, 128%-180% higher CNRD, and 22% lower median dose (38.4 versus 49.2 mGy; all p<.001). All three primary endpoints remained significant in the ≤1-day subset (n=60). Artifact scores did not differ significantly. Exploratory contrast-enhanced and angiographic analyses favored PCCT for selected parenchymal and small-vessel metrics, including perforator visualization (median 3 versus 2; p<.001), but these small subgroups were preliminary and did not strengthen the primary conclusions.

PCCT was associated with higher image quality, pathology conspicuity, GM/WM CNR and CNRD, and lower dose in routine non-contrast brain CT. Smaller protocol groups were exploratory. Because scanner generations, vendors, protocols, and reconstructions were unmatched, the differences cannot be attributed solely to detector technology.