A pediatric anthropomorphic phantom (ATOM Phantom; CIRS, Norfolk, Virginia, USA) representing an average newborn was used in this study. The expected weight and height of the neonate were 3.5 kg and 51.0 cm, respectively. A pediatric phantom is made of radiologically equivalent tissue material with internal structures, such as an artificial skeleton, lungs, and soft tissue, to accurately simulate clinical exposure.
Positioning of the subject
In the conventional method, the pediatric phantom is positioned at a height of 0 cm above the table. In the air-gap method, the pediatric phantom was positioned at a height of 15 cm above the table using 15-cm-thick polystyrene with low X-ray absorption (Figure 1).
A 64-detector row CT scanner (Lightspeed VCT; GE Healthcare, Milwaukee, WI, USA) was used to perform the helical scan from the lung apex to the subdiaphragm. The scan range of the neonatal phantom was 100 mm, and scan parameters were as follows: helical mode, helical pitch 0.986, beam width of 40.0 mm, section thickness of 5.0 mm, gantry rotation time of 0.4 s, 64 × 0.625-mm detector collimation, a small scan field of view setting of 100 mm, a full-mode matrix size 512 × 512 mode, and a standard reconstruction kernel. The tube voltage was 80 kVp, the tube current was fixed at 50 mA, which is commonly used in clinical practice, and CT images were acquired using a filtered back-projection algorithm under a standard kernel/filter.
Dosimeters and dose measurement and computed tomography dose index
A real-time skin dosimeter (RD - 1000; Trek Corporation, Kanagawa, Japan) was placed in the scanner gantry, and the RD - 1000 dosimeter was positioned at the center of the phantom, on the dorsal surface of the body, and on the left and right mammary glands (Figure 2) . The pediatric anthropomorphic phantom was then scanned 10 times for each method, and the measured dose values of the RD - 1000 were compared for both the conventional and air-gap methods. We then measured the CT dose index volume (CTDIvol) values on the CT console.
Comparison of radiation doses from 10-cm ionization chamber, computed tomography dose index volume, and RD-1000
Because it was necessary to confirm the accuracy of the CTDIvol of the console-displayed dose, we compared the measured CTDIvol of the console-displayed dose for the CT system and the 10-cm ionization chamber. The CTDIvol showed good linearity of the reference dose with the 10-cm ionization chamber. Because it was necessary to confirm the traceability of the RD-1000, we compared the doses from the 10-cm ionization chamber and the RD-1000 using a general X-ray radiography system. The RD-1000 demonstrated good linearity of the reference dose with the ionization chamber.
Measurement of image noise
The image noise [standard deviation (SD) of CT values] was measured at the center of the phantom at the level of the mammary glands area within a 10.0-mm diameter region of interest. A radiologist with 26 years of experience performed the measurements. For each scan, image noise was measured five times at a total of five points at the margins and at the center. The mean of the image noise was calculated and the image noise was compared for each protocol.
The Mann-Whitney U test was used to compare the measured dose values, image noise, and CTDIvol measurements; differences with a P-value < 0.05 were considered statistically significant. Statistical analyses were performed using the free statistical software platform R (version 3.0.2; R Project for Statistical Computing, http://www.rproject.org/).