Developing optimized CT scan protocols : phantom measurements of image quality

Abstract

Purpose: The increasing frequency of computerized tomography (CT) examinations is well documented, leading to concern about potential radiation risks for patients. However, the consequences of not performing the CT examination and missing injuries and disease are potentially serious, impacting upon correct patient management. The ALARA principle of dose optimization must be employed for all justified CT examinations. Dose indicators displayed on the CT console as either CT dose index (CTDI) and/or dose length product (DLP), are used to indicate dose and can quantify improvements achieved through optimization. Key scan parameters contributing to dose have been identified in previous literature and in previous work by our group. The aim of this study was to optimize the scan parameters of mA; kV and pitch, whilst maintaining image quality and reducing dose. This research was conducted using psychophysical image quality measurements on a CT quality assurance (QA) phantom establishing the impact of dose optimization on image quality parameters.

Method: Current CT scan parameters for head (posterior fossa and cerebrum), abdomen and chest examinations were collected from 57% of CT suites available nationally in Malta (n ¼ 4). Current scan protocols were used to image a Catphan 600 CT QA phantom whereby image quality was assessed. Each scan parameter: mA; kV and pitch were systematically reduced until the contrast resolution (CR), spatial resolution (SR) and noise were significantly lowered. The Catphan 600 images, produced by the range of protocols, were evaluated by 2 expert observers assessing CR, SR and noise. The protocol considered as the optimization threshold was just above the setting that resulted in a significant reduction in CR and noise but not affecting SR at the 95% confidence interval.

Results: The limit of optimization threshold was determined for each CT suite. Employing optimized parameters, CTDI and DLP were both significantly reduced (p 0.001) by 0e49% and 0e51% respectively. Below the optimization threshold, CR and noise were significantly affected (p 0.05) but not SR (p 0.083). The threshold value is the value below which CR and noise were affected.

Conclusion: Optimization was effected through the establishment of the limits at which image quality is detrimentally reduced by dose reduction. This work established the mA, kV and pitch changes necessary to achieve the optimization thresholds, reducing the dose and the limit at which psychophysical image quality is significantly decreased. Further work will evaluate these settings in the clinical environment.