Impact of Spatial Resolution on CT Imaging of Middle Ear Prostheses: Comparison of Photon-Counting and Energy-Integrating Detectors.

BACKGROUND AND PURPOSE: CT imaging of submillimeter middle ear prostheses-vital for postsurgery treatment-requires high spatial resolution. This study evaluated the impact of spatial resolution for the visualization of prostheses using photon-counting detector (PCD) CT, energy-integrating detector (EID) CT, and EID-CT with attenuating comb filters.

MATERIALS AND METHODS: Twenty-one middle ear prostheses of different makes were placed within a head phantom and scanned on a PCD-CT (NAEOTOM Alpha), 2 dual-source CTs with comb filters (Force and Flash), and a dual-layer CT without comb filter (7500). Scanning and reconstruction followed clinical temporal bone protocols for each scanner, with a matched 120 kV and 80 mm FOV; sharp kernels/algorithms of Qr89, Ur77, Ur77, and YD; and minimal allowable slice thickness of 0.2, 0.4, 0.5, and 0.67 mm, respectively. Qualitative assessments were performed, including visual inspection, volume rendering, a blind reader study, and quantitative assessments using line profiles, and full width at half maximum (FWHM). One prosthesis was implanted in human cadaver head and scanned on a PCD-CT (NAEOTOM Alpha), a third-generation dual-source (Force), and a dual-layer CT (7500) following the same phantom study protocols. Image quality was compared among the scanners.

RESULTS: PCD-CT provided images with the highest spatial resolution, allowing clinically relevant assessment of prosthesis position. EID-CT scanners with comb filters provided good results, with noticeable loss of detail, while EID-CT without a comb filter showed substantial resolution loss. These findings were consistent across readers (Kendall W = 0.89; P < .001) and validated by the line profiles and FWHM measurements. PCD-CT showed a 49.5% improvement over EID-CT without comb filter. Among scanners with ultra-high-resolution capabilities, PCD-CT had lower radiation dose (CTDIvol: 31.5 versus 46.7 and 51.4 mGy) and lower noise (81.3 versus 93.0 and 274.8 Hounsfield Units (HU) compared with 2 EID-CTs with comb filters. In the cadaver study, PCD-CT provided higher resolution and lower noise than both EID-CT scanners with and without comb filters (67.6 versus 118.6 and 50.1 HU).

CONCLUSIONS: PCD-CT provided dose-efficient, accurate evaluation of ossicular prostheses, without the degradation of submillimeter structures, while maintaining lower dose and noise.