Abstract
OBJECTIVES: This study evaluates the effectiveness of spectral filtration-specifically Tin and SilverBeam filters-in achieving ultra-low radiation doses in pediatric computed tomography (CT) imaging for craniosynostosis diagnosis. We investigate whether these filters can reduce radiation to levels comparable to or below those of standard four-view skull x-rays, while maintaining diagnostic accuracy. Unlike previous research focused broadly on dose reduction, this study highlights the potential of Tin and SilverBeam filtration as a promising solution. METHODS: CT images were acquired using a pediatric head fracture phantom, with sutures simulating craniosynostosis, on two scanners with different spectral filters. The CTDIvol was reduced by varying percentages from standard pediatric protocols using Tin and SilverBeam filters. Image quality and radiation dose were quantitatively assessed, while two radiologists performed qualitative evaluations. RESULTS: Dose comparisons showed that images with at least a 89.2% reduction in CTDIvol using the Tin filter and a 91.4% reduction when using the SilverBeam filter resulted in lower doses than a standard four-view skull x-ray (0.09 mSv). The greatest dose reduction (up to 161%) occurred with SilverBeam at a 99.9% reduction in CTDIvol. While significant pixel intensity changes were observed at the sutures, spatial resolution decreased at lower dose settings. Two pediatric radiologists observed clear skull and orbit outlines under clinical conditions, but coronal sutures became undetectable near 90% CTDIvol reduction. CONCLUSION: Spectral filtration ultra-low-dose CT significantly reduces radiation exposure compared to four-view skull x-rays, showing promise for diagnosing craniosynostosis. Further patient-based studies are needed to validate diagnostic efficacy. ADVANCES IN KNOWLEDGE: This study is the first to assess spectral filtration in ultra-low-dose pediatric head CT, highlighting significant dose reductions without compromising diagnostic quality, and stressing the need for further validation.