All imaging data of patients under the age of 16 years which were scanned on an abdominal protocol in 2016 in the Leuven University Hospital were gathered. All scans were taken on one of two available Siemens Somatom Definition Flash-scanners (Siemens Healthcare, Forchheim, Germany).
Gathered data included the used tube current (mA), the used reference tube voltage (kV), the Volume Computed Tomography Dose Index (CTDIvol), patients' length and weight and clinical indication.
A difference is made between child scan protocols versus adult scan protocols, and larger children can be scanned on adult protocols. These protocols correspond with a difference in tube current modulation by the CareDOSE 4D system (Siemens Healthcare, Forchheim, Germany) of respectively ‘weak’ versus ‘average’.
Image quality was subjectively scored by a last year medical student.
1. Doses were compared to the guidelines as defined in the pediatric imaging Dose Reference Levels (PiDRL). The PiDRL-guidelines use CTDIvol and define DRL’s for the following subcategories: 5-15kg (<3.5mGy), 15-30kg (<5.4mGy), 30-50kg (<7.3mGy) and 50-80kg (<13mGy).[5]
Weight was available for 35 of 38 patients. The average weight was 27.79 kg, with the smallest patient weighing 7.34 kg, and the heaviest patient weighing 54.4 kg. The standard deviation was 15.14 kg. One scan was identified with a dose larger than the PiDRL. This patient weighed 48.5kg and had a CTDIvol of 7.52mGy. (Figure 1)
2. To study if radiation doses did indeed lower with patient size, weights and CTDIvol of patients were analysed. Linear regression analysis showed a correlation of R= 0.859 with a statistical significance of p=3.86 E-11 (p<0.05) (Figure 2).
3. Image quality was investigated for a correlation with radiation dose.
All scans were blinded and scored by a last year medical student on a single picture archiving and communication system (PACS). A visual grading scale from 1 to 4 was used. There were no non-assessable CT examinations (score= 1), seven scans were determined to be of ‘inferior quality’ (score= 2), twenty-five scans were ‘adequate’ (score= 3) and six scans were scored ‘excellent’ (score= 4).
A moderate correlation was found between image quality and radiation dose with R= 0.482, and p-value= 0.0025 (p<0.05).
All seven inferior protocols were scanned with CAREDose4D setting on ‘weak’. Of the six excellent scans, three were scanned witch CAREDose4D on ‘average’. (Figure 3)