Optimization of reconstruction strategies for head and neck computed tomography angiography using photon-counting...

Head and neck computed tomography angiography (CTA) requires high spatial resolution while maintaining acceptable image noise. This study aimed to compare image quality between photon-counting CT (PCCT) and energy-integrating detector CT (EID-CT) and to evaluate PCCT reconstruction strategies for specific clinical imaging tasks.

This retrospective single-center study included 113 head and neck CTA examinations, comprising 60 PCCT and 53 energy-integrating detector CT (EID-CT) examinations. EID-CT images were reconstructed at 1.0 mm using the Bv40 kernel. PCCT datasets were reconstructed using the following strategies: 1.0-mm Bv40; 0.4-mm Bv44, Bv48, Bv56, and Bv60; and 0.2-mm Bv64 and Bv72. Quantitative image quality metrics included vascular attenuation, image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and vessel sharpness. Two neuroradiologists independently assessed qualitative image quality. Blooming artifacts were evaluated in the calcified plaque and stent subgroups.

PCCT showed lower CTDIvol (23.3 ± 3.0 vs 26.2 ± 4.4 mGy), dose-length product (795.6 ± 113.0 vs 891.6 ± 159.3 mGy·cm), and effective dose (2.5 ± 0.4 vs 2.8 ± 0.5 mSv) than EID-CT (all p < 0.01). Thinner and sharper PCCT reconstructions improved vessel sharpness but increased image noise and reduced SNR and CNR. Subjective image quality remained high, with substantial inter-reader agreement (mean weighted κ, 0.759-0.819). Calcified-plaque blooming decreased from 0.35 ± 0.08 with 1.0-mm Bv40 to 0.22 ± 0.04 with 0.2-mm Bv72 (p < 0.01).

Under the evaluated clinical protocols, the PCCT group showed lower radiation exposure while maintaining high image quality compared with the EID-CT group. Different PCCT reconstruction strategies had distinct effects on image quality, suggesting that reconstruction selection should be tailored to specific clinical tasks by balancing spatial resolution and image noise.