Hisanobu Koyama MD, PhD, Presenter: Nothing to Disclose

Yoshiharu Ohno MD, PhD, Abstract Co-Author: Research Grant, Toshiba Corporation Research Grant, Koninklijke Philips NV Research Grant, Bayer AG Research Grant, DAIICHI SANKYO Group Research Grant, Eisai Co, Ltd Research Grant, Terumo Corporation Research Grant, Fuji Yakuhin Co, Ltd Research Grant, FUJIFILM Holdings Corporation Research Grant, Guerbet SA

Shinichiro Seki, Abstract Co-Author: Nothing to Disclose

Mizuho Nishio MD, PhD, Abstract Co-Author: Research Grant, Toshiba Corporation

Hiroyasu Inokawa, Abstract Co-Author: Employee, Toshiba Corporation

Naoki Sugihara MENG, Abstract Co-Author: Employee, Toshiba Corporation

Noriyuki Negi RT, Abstract Co-Author: Nothing to Disclose

Tohru Murakami, Abstract Co-Author: Nothing to Disclose

Takeshi Yoshikawa MD, Abstract Co-Author: Research Grant, Toshiba Corporation

Sumiaki Matsumoto MD, PhD, Abstract Co-Author: Research Grant, Toshiba Corporation

Kazuro Sugimura MD, PhD, Abstract Co-Author: Research Grant, Toshiba Corporation Research Grant, Koninklijke Philips NV Research Grant, Bayer AG Research Grant, Eisai Co, Ltd Research Grant, DAIICHI SANKYO Group

To compare lung nodule detection capability of ultra-low-dose and low-dose CTs among newly developed full iterative reconstruction techniques (FIR), clinically available iterative reconstruction (i.e. adaptive iterative dose reduction using three dimensional processing <AIDR 3D>) and filter back projection (FBP) techniques in chest phantom study.

A chest CT phantom including simulated GGOs (-800HU) and part solid nodules (-630HU), was scanned on area-detector CT at standard-dose CT (SDCT: 270mA), low-dose CT (LDCT: 50mA) and ultra-low-dose CT (ULDCT: 10mA) protocols. Then, all CT data sets were reconstructed with FIR, AIDR 3D and/ or FBP. For quantitative image quality assessment, image noise at each protocol was assessed by ROI measurements. To determine the capability of nodule identification on each protocol, two chest radiologists independently evaluated lesion conspicuity at each nodule by means of 5-point scoring system, and final scores were made by consensus of two readers. Image noise was compared each other by Tukey's HSD test at each tube current. Then, ROC analyses were performed to compare identification capability among all techniques at each tube current, between SDCT and each LDCT, and between SDCT and each ULDCT.

Image noises of FBP were significantly higher than that of others at each tube current (p<0.05). In addition, image noise of FIR was the lowest at both tube currents. When compared identification capability, area under the curves (Az) of LDCT and ULDCT reconstructed with FIR (LDCT: Az=0.94, ULDCT: Az=0.90) and those with AIDR 3D (LDCT: Az=0.94, ULDCT: Az=0.90) were significantly higher than those with FBP (LDCT: Az=0.91, p<0.05; ULDCT: Az=0.70, p<0.05). When compared with SDCT (Az=0.95), identification capability of ULDCT with each method was significantly lower than that of SDCT (p<0.05) in this setting.

Newly developed FIR algorithm as well as AIDR 3D is useful for LDCT and ULDCT, and can improve image quality and nodule identification as compared with FBP at each tube current level.

On low- and ultra-low-dose CT, newly developed full iterative reconstruction algorithm as well as commercially available iterative reconstruction technique is useful than filter back projection for improving image quality and nodule identification.

Koyama, H, Ohno, Y, Seki, S, Nishio, M, Inokawa, H, Sugihara, N, Negi, N, Murakami, T, Yoshikawa, T, Matsumoto, S, Sugimura, K, Comparisons of Lung Nodule Detection Capability on Ultra-low and Low-dose CTs among Newly Developed Full Iterative Reconstruction, Clinically Available Adaptive Iterative Dose Reduction 3D and Filter Back Projection Techniques in Chest Phantom Study.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14007768.html