Author: Samuel L. Brady, Kevin Chen, Joseph G. Meier 👨🔬
Affiliation: University of Cincinnati, Cincinnati Childrens Hospital Med Ctr 🌍
Purpose: Conebeam-CT (CBCT) acquisition protocols typically do not distinguish between adults and pediatrics. In collaboration with a fluoroscopically-guided interventional (FGI) manufacturer, new, dose-reduced protocols were developed specifically for pediatric patients. This study investigates the impact on image quality as a function of radiation dose and patient size.
Methods: A Corgi phantom was used to assess image quality metrics such as: sensitometry accuracy, image noise, noise power spectrum (NPS), task transfer function (TTF), and non-prewhitening filter detectability index (dNPWe'). Additionally, detection confidence was assessed by calculating the area under the curve (AUC). The phantom was 13.5 cm in diameter and was scanned with annuli to represent additional effective diameters of 16 cm and 20 cm. Radiation dose was assessed using MOSFETs placed within three anthropomorphic phantoms representing a 1-year-old (13 cm diameter), 5-year-old (15 cm thorax to 17 cm abdomen diameter), and a teenager (20 cm to 22 cm diameter). Standard-of-care roll and propeller CBCT acquisitions were compared to reduced-dose protocols.
Results: Dose-reduced roll protocols resulted in a 44%-70% reduction in thoracic dose and a 41%-70% reduction in abdominal dose. For propeller, thoracic dose was reduced by 45%, and abdominal dose by 35%-46%. Sensitometry values differed by only 1%-4% from the standard protocol. At 50% contrast, the TTF was similar between standard and reduced-dose protocols: 0.8 lp/cm for all phantom sizes in propeller, and 0.8, 0.7, and 0.6 lp/cm for small, medium, and large phantoms in roll. Noise variance doubled for dose-reduced protocols, but the NPS peak frequency remained consistent at 0.6 lp/cm across all phantom sizes. Detection confidence for vessels < 2 mm was reduced by 5% for dose-reduced CBCT.
Conclusion: Dose-reduced protocols were reduced by 33%-70%. Despite increased noise variance, image quality (spatial resolution, sensitometry, noise texture) remained stable, with a minimal reduction in detection confidence for small vessels.