INTRODUCTION Rapid adoption of ultra-low-dose CT protocols (∼1-2 mGy) used in paediatric cardiac CT examinations challenges current quality control (QC). The IEC 61223-3-5 constancy criterion for CTDIvol display accuracy ("±20% or ±1 mGy") permits up to ±100% relative error at 1.0 mGy, undermining patient dose management and early detection of equipment anomalies. This study evaluates ultra-low-dose display accuracy and proposes revised QC benchmarks. METHODS CTDIvol accuracy was evaluated across two state-of-the-art CT systems using a cylindrical polymethyl methacrylate (PMMA) phantom with a 32 cm-diameter and a 15 cm-length and a calibrated semiconductor detector. Measurements ranged from 10 to 0.5 mGy (120, 100, and 80 kV), utilizing synchronized and asynchronous rotational exposure methods to ensure geometric reproducibility. RESULTS Measured doses remained within a ±20% relative tolerance across a tested range from 0.5 mGy to 10.3 mGy, proving precise control is achievable in the ultra-low-dose range. However, system-specific discrepancies were identified, with manufacturers showing distinct overestimation or underestimation tendencies. Crucially, applying the ±1 mGy absolute tolerance below 5.0 mGy permitted exponentially increasing relative errors, effectively nullifying fundamental QC objectives. CONCLUSION The ±1 mGy absolute constancy criterion loses effectiveness in the ultra-low-dose regime, creating a critical latent risk for undetected equipment degradation. International QC standards urgently require revision to ensure reliable dose monitoring. IMPLICATIONS FOR PRACTICE We propose establishing a "low-dose boundary value" at 5.0 mGy. Above this threshold, the existing ±1 mGy criterion applies; below it, a uniform ±20% relative tolerance is strictly enforced. This provides a robust, backward-compatible QC framework for low-dose CT, highlighting the need to recognize manufacturer-dependent display uncertainties during protocol optimization.
INTRODUCTION Lower-extremity CT angiography (CTA) is widely used for the assessment of peripheral arterial disease. Reducing iodinated contrast medium is desirable, particularly in patients with impaired renal function. This study evaluated the impact of 40-keV virtual monoenergetic images (VMI) on image quality in low-contrast lower-extremity spectral CTA. METHODS In this retrospective single-center study, 60 patients who underwent lower-extremity spectral CTA with 40 mL of iodinated contrast medium were included. Conventional reconstructions and 40-keV VMI were compared regarding vascular attenuation, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), iodine concentration, and subjective image quality across predefined arterial segments. Qualitative image quality was independently assessed by two readers using a 5-point Likert scale. RESULTS Compared with conventional reconstructions, 40-keV VMI significantly increased vascular attenuation, SNR, and CNR across all arterial segments (all p < 0.001). Subjective ratings for vessel contrast, vessel sharpness, overall image quality, and diagnostic confidence were significantly higher with VMI for both readers across all evaluated arterial segments (all p < 0.05). Iodine concentration decreased progressively from proximal to distal arterial segments, whereas the CNR advantage of VMI remained preserved throughout the arterial tree. CONCLUSION Forty-keV virtual monoenergetic reconstruction significantly improves quantitative and qualitative image-quality parameters in low-contrast lower-extremity spectral CTA compared with conventional reconstructions. These findings suggest that low-keV VMI may improve image-quality characteristics in reduced-contrast CTA protocols, particularly in distal arterial segments. IMPLICATIONS FOR PRACTICE Low-keV virtual monoenergetic reconstruction may support reduced-contrast lower-extremity CTA protocols by improving image quality without increasing contrast medium volume. This approach may be particularly valuable for patients in whom minimizing iodinated contrast medium exposure is desirable.
