Masaki Katsura MD, Presenter: Nothing to Disclose

Izuru Matsuda MD, Abstract Co-Author: Nothing to Disclose

Masaaki Akahane MD, Abstract Co-Author: Nothing to Disclose

Koichiro Yasaka MD, Abstract Co-Author: Nothing to Disclose

Shouhei Hanaoka MD, Abstract Co-Author: Nothing to Disclose

Hiroyuki Akai MD, Abstract Co-Author: Nothing to Disclose

Jiro Sato MD, Abstract Co-Author: Nothing to Disclose

Akira Kunimatsu MD, Abstract Co-Author: Speakers Bureau, Eisai Co, Ltd Speakers Bureau, General Electric Company

Kuni Ohtomo MD, Abstract Co-Author: Research Grant, Bayer AG Research Grant, DAIICHI SANKYO Group

To evaluate the effects of pure and hybrid iterative reconstruction techniques on computer-aided detection (CAD) system of pulmonary nodules using reference-, low-, and ultralow-dose chest CT.

Fifty-nine patients(mean[SD] age, 64.7[13.4] years) gave informed consent to undergo reference-, low-, and ultralow-dose chest with 64-row multidetector CT. The reference- and low-dose CT involved the use of automatic tube current modulation with fixed noise indices(31.5&70.44 at 0.625mm, respectively) and the ultralow-dose CT was acquired with a fixed tube current-time product of 5mAs. Each 0.625-mm-thick image was reconstructed with filter-back projection(FBP), 50% adaptive statistical iterative reconstruction(ASIR)-FBP blending, and model-based iterative reconstruction(MBIR). CAD output was compared with the results of the reference standard which was established using a consensus panel of 2 radiologists, who identified 84 non-calcified nodules with diameters of 4mm or greater on reference-dose ASIR. Sensitivity in nodule detection between different radiation dose and reconstruction techniques was assessed using Cochran's Q test with post hoc McNemar test and Holm correction for multiple comparisons. The mean number of false-positive findings per examination was assessed using Friedman test with post hoc testing using Wilcoxon signed-rank test and Holm correction.

Compared with reference-dose CT(307.7±178.1mGy-cm), there was a 78.6% decrease in dose–length product with low-dose CT(66.0±50.8mGy-cm), and a 95.3% decrease with ultralow-dose CT(14.5±1.1mGy-cm). CAD sensitivity for nodules at ultralow-dose MBIR(26.2%) was equivalent(p=0.52-1.00) to those at reference-(FBP,23.8%;ASIR,26.2%,MBIR;26.2%) and low-dose CT(FBP,26.2%;ASIR,25.0%,MBIR;25.0%). Mean number of false-positive findings per examination tended to increase with the use of MBIR, particularly at low-(1.19) and ultralow-dose CT(1.32), although there were no significant differences among reconstruction techniques at both dose levels(p=0.08-1.00).

CAD sensitivity for pulmonary nodules at ultra-dose MBIR is almost equal to that at reference-dose CT. MBIR can improve CAD performance on dose-reduced chest CT despite increased false-positive findings.

With the use of MBIR, CAD system on ultralow-dose CT may play an equivalent role to that on regular-dose CT in helping radiologists detect pulmonary nodules.

Katsura, M, Matsuda, I, Akahane, M, Yasaka, K, Hanaoka, S, Akai, H, Sato, J, Kunimatsu, A, Ohtomo, K, The Effects of Pure and Hybrid Iterative Reconstruction Techniques on Computer-aided Detection (CAD) System of Pulmonary Nodules in Dose-reduced CT Scans.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13018496.html