Lung nodule detection performance in five observers on computed tomography (CT) with adaptive iterative dose reduction using three-dimensional processing (AIDR 3D) in a Japanese multicenter study: Comparison between ultra-low-dose CT and low-dose CT by receiver-operating characteristic analysis

• Published in: European journal of radiology Vol. 84; no. 7; pp. 1401 - 1412
• Main Authors: Nagatani, Yukihiro, Takahashi, Masashi, Murata, Kiyoshi, Ikeda, Mitsuru, Yamashiro, Tsuneo, Miyara, Tetsuhiro, Koyama, Hisanobu, Koyama, Mitsuhiro, Sato, Yukihisa, Moriya, Hiroshi, Noma, Satoshi, Tomiyama, Noriyuki, Ohno, Yoshiharu, Murayama, Sadayuki
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier Ireland Ltd 01.07.2015
• Subjects: Adaptive iterative dose reduction; Aged; Aged, 80 and over; Body Mass Index; Chest; Clinical Competence - statistics & numerical data; Computed tomography; Female; Humans; Image quality; Imaging, Three-Dimensional - methods; Japan; Lung - diagnostic imaging; Lung Neoplasms - diagnostic imaging; Male; Middle Aged; Observer Variation; Radiation Dosage; Radiographic Image Interpretation, Computer-Assisted - methods; Radiography, Thoracic - methods; Radiology; ROC Curve; Tomography, X-Ray Computed - methods; Ultra-low dose scanning; Image quality; Chest; Computed tomography; Adaptive iterative dose reduction; Ultra-low dose scanning
• ISSN: 0720-048X; 1872-7727; 1872-7727
• DOI: 10.1016/j.ejrad.2015.03.012

•Using AIDR 3D, ULDCT showed comparable LND of solid nodules to LDCT.•Using AIDR 3D, LND of smaller GGN in ULDCT was inferior to that in LDCT.•Effective dose in ULDCT was about only twice of that in chest X-ray.•BMI values in study population were mostly in the normal range body habitus. To compare lung nodule detection performance (LNDP) in computed tomography (CT) with adaptive iterative dose reduction using three dimensional processing (AIDR3D) between ultra-low dose CT (ULDCT) and low dose CT (LDCT). This was part of the Area-detector Computed Tomography for the Investigation of Thoracic Diseases (ACTIve) Study, a multicenter research project being conducted in Japan. Institutional Review Board approved this study and informed consent was obtained. Eighty-three subjects (body mass index, 23.3±3.2) underwent chest CT at 6 institutions using identical scanners and protocols. In a single visit, each subject was scanned using different tube currents: 240, 120 and 20mA (3.52, 1.74 and 0.29mSv, respectively). Axial CT images with 2-mm thickness/increment were reconstructed using AIDR3D. Standard of reference (SOR) was determined based on CT images at 240mA by consensus reading of 2 board-certificated radiologists as to the presence of lung nodules with the longest diameter (LD) of more than 3mm. Another 5 radiologists independently assessed and recorded presence/absence of lung nodules and their locations by continuously-distributed rating in CT images at 20mA (ULDCT) and 120mA (LDCT). Receiver-operating characteristic (ROC) analysis was used to evaluate LNDP of both methods in total and also in subgroups classified by LD (>4, 6 and 8mm) and nodular characteristics (solid and ground glass nodules). For SOR, 161 solid and 60 ground glass nodules were identified. No significant difference in LNDP for entire solid nodules was demonstrated between both methods, as area under ROC curve (AUC) was 0.844±0.017 in ULDCT and 0.876±0.026 in LDCT (p=0.057). For ground glass nodules with LD 8mm or more, LNDP was similar between both methods, as AUC 0.899±0.038 in ULDCT and 0.941±0.030 in LDCT. (p=0.144). ULDCT using AIDR3D with an equivalent radiation dose to chest x-ray could have comparable LNDP to LDCT with AIDR3D except for smaller ground glass nodules in cases with normal range body habitus.