Radiation dose optimisation for conventional imaging in infants and newborns using automatic dose management software: an application of the new 2013/59 EURATOM directive

Abstract

Objective: The new 2013/59 EURATOM Directive (ED) demands dosimetric optimisation procedures without undue delay. The aim of this study was to optimise paediatric conventional radiology examinations applying the ED without compromising the clinical diagnosis.

Methods: Automatic dose management software (ADMS) was used to analyse 2678 studies of children from birth to 5 years of age, obtaining local diagnostic reference levels (DRLs) in terms of entrance surface air kerma. Given local DRL for infants and chest examinations exceeded the European Commission (EC) DRL, an optimisation was performed decreasing the kVp and applying the automatic control exposure. To assess the image quality, an analysis of high-contrast resolution (HCSR), signal-to-noise ratio (SNR) and figure of merit (FOM) was performed, as well as a blind test based on the generalised estimating equations method.

Results: For newborns and chest examinations, the local DRL exceeded the EC DRL by 113%. After the optimisation, a reduction of 54% was obtained. No significant differences were found in the image quality blind test. A decrease in SNR (-37%) and HCSR (-68%), and an increase in FOM (42%), was observed.

Conclusion: ADMS allows the fast calculation of local DRLs and the performance of optimisation procedures in babies without delay. However, physical and clinical analyses of image quality remain to be needed to ensure the diagnostic integrity after the optimisation process. Advances in knowledge: ADMS are useful to detect radiation protection problems and to perform optimisation procedures in paediatric conventional imaging without undue delay, as ED requires.

Figure 1.

TOR CDR image used to obtain the physical image metrics for a 4 cm PMMA thickness. ROIs in the first low-contrast circle, background (BG) and 7th group bar pattern ROI (ROI 7th) are shown. BG, background; PMMA, polymethylmethacrylate; ROI, region-of-interest.

Figure 2.

K_a,e distributions before (red) and after (blue) the action in the room for the Chest PA/AP examinations in the age ranges considered: (a) (0, 3) months of age, (b) (3, 6) months of age, (c) (6, 12) months of age. The red lines represent the EC DRLs. AP, anteroposterior; DRLs, diagnostic reference levels; EC, European Commission; PA, posteroanterior.

Figure 3.

Physical image quality depending on slab thickness in terms of SNR (a) and HCSR (b), obtained before and after the in-room action. HCSR, high-contrast resolution; SNR, signal-to-noise ratio.

Figure 4.

Relationship between physical image quality and dose to the patient, obtained in terms of FOM before and after the in-room action. FOM, figure of merit.

Figure 5.

(a) Two images of patients (7-month-old to the left, 3-month-old to the right) obtained after the action in the room with no visualisation of the costophrenic angles (score 0) due to pulmonary diseases (red arrows). All the radiologists scored these images in the same manner; (b) Two images of patients obtained after the action in the room with the poorest evaluation of noise (score 0). The image on the left (3-month-old) was scored with null value by all the radiologists. The image on the right (4-month-old) was scored with null value only by radiologist 3.

Figure 6.

Clinical image quality analysis results corresponding to the general evaluation. Optimum contrast (a), noise (b), well-defined edges (c) and optimum general acceptability (d) evaluations are shown before and after the action in the room (Total columns), including the evaluation performed by each radiologist separately (R1, R2 and R3). All the p-values are shown.