Enhanced Soft Tissue Contrast In Neonatal Brain Imaging Using Deep Silicon Photon Counting CT (dSi-PCCT): A Custom Phantom Study
Abstract
Purpose
To evaluate image quality in neonatal brain CT using a custom phantom designed to replicate clinically relevant skull and soft tissue attenuation, enabling assessment of contrast-to-noise ratio (CNR) for improved diagnostic confidence in pediatric neuroimaging.
Methods
A neonatal brain phantom was developed with skull-mimicking material and vials containing NaCl solutions simulating cerebrospinal fluid (CSF) and neonatal brain tissue with attenuation differences of 10-20HU. Scans were acquired at clinically relevant CTDIvol of 18 mGy on deep silicon-based photon-counting CT (dSi-PCCT, 65 keV) and energy-integrated detector CT (EID CT, 80 and 100 kVp) using pediatric SFOV settings. Images were reconstructed with medium strength of denoising, matched-resolution soft tissue kernel, 1.25 mm slice thickness, and 512 × 512 matrix. CNR between the CSF and brain tissue surrogates was measured.
Results
dSi‑PCCT demonstrated CNR improvements of 15.5% and 24.0% relative to EID CT at 80 kVp and 100 kVp, respectively. Importantly, even though dSi‑PCCT is currently limited to 120 kVp acquisition, its performance remained superior when compared to the lower‑kVp EID acquisitions. The broader photon‑energy distribution associated with higher kVp provided enhanced spectral information for material decomposition, contributing to additional CNR gains. Monochromatic 65 keV reconstructions further improved visualization of low‑contrast structures compared with low-kVp EID CT, aiding detection of subtle neonatal intracranial abnormalities.
Conclusion
Precise evaluation of neonatal CT image quality is essential, as small attenuation differences may reflect early brain pathology. The custom phantom enabled realistic and reproducible assessment of CNR under pediatric imaging conditions. dSi‑PCCT consistently outperformed conventional low‑kVp EID CT acquisitions, underscoring its potential to improve diagnostic confidence and reduce repeat scans in vulnerable neonatal patients.