The application of dose-reduction simulation in neonatal head CT scan / 中华放射学杂志

ABSTRACT

Objective To determine the effects of dose reduction on multi-slice spiral CT(MSCT) of neonatal head and assess the lowest possible radiation for acceptable clinical images.Methods Fifty-seven newborns suspected intracranial hemorrhage were entered into the study and underwent MSCT scans.Original images at three anatomic levels (posterior fossa, basal ganglia, centrum semiovale) were collected and synthetic noise was added so as to simulate dose reductions of 25%, 40%, 50% and 70%, respectively by using the noise addition tool.A total of 855 image data sets were obtained for the 57 patients.Original and simulated dose-reduction scan images were analyzed.Image noise and image quality were assessed by two independent experienced pediatric radiologists using diagnostic acceptability score, subjective image noise score on a 5-peint scale and objective noise index.Image noise was measured by respectively placing region of interest (ROI) at cerebellum, thalamus and corona radiata of 3 different slices.And the noise index and mean value was calculated.The degree of inter-observer concordance was determined by Kappa statistical analysis.The Spearman statistical correlations between the noise index and diagnostic acceptability score were performed.Results On the images of original dose and simulated dose reductions of 25%, 40%, 50% and 70%, the diagnostic acceptability was 4.47±0.51, 3.96±0.33, 3.21±0.45, 2.92±0.32, and 1.85±0.57, respectively,the subjective image scores were 1.62±0.48, 1.99±0.48, 2.76±0.81, 3.19±0.67, and 4.27±0.54, respectively, the noise index were 1.90±0.19, 2.17±0.20.2.44± 0.25, 2.68±0.28, and 3.37±0.39, respectively.The two radiologists had good intembserver agreement for diagnostic acceptability (K=0.860, P=0.017) and for image noise scoff ng (K=0.630, P=0.022).There was significant statistical correlation between image noise index and diagnostic acceptability (r= 0.826,P=0.001).At 40% dose reduction to the standard protocol, the noise index was 2.44 and the image quality score was 3.21 which were considered clinically acceptable.Conclusion The study revealed that acceptable imagos could be obtained with mean noise index of 2.44 and 40% dose reduction.