Model-Based Iterative Reconstruction Technique for Ultralow-Dose Chest CT

• DOI: 10.1097/rli.0b013e31827efc3a
• Publication Date: 2013-01-23T11:51:32Z
• Authors: Masaki Katsura; Izuru Matsuda; Masaaki Akahane; Koichiro Yasaka; Shohei Hanaoka; Hiroyuki Akai; Jiro Sato; Akira Kunimatsu; Kuni Ohtomo
• Keywords: McNemar's test

The purpose of this study was to evaluate whether model-based iterative reconstruction (MBIR) enables dose reduction over adaptive (ASIR) while maintaining diagnostic performance.In institutional review board-approved and Health Insurance Portability Accountability Act-compliant study, 59 patients (mean [SD] age, 64.7 [13.4] years) gave informed consent undergo reference-, low-, ultralow-dose chest computed tomography (CT) with 64-row multidetector CT. reference- low-dose CT involved the use automatic tube current modulation fixed noise indices (31.5 70.44 at 0.625 mm, respectively) were reconstructed 50% ASIR-filtered back projection blending. acquired a current-time product 5 mA s MBIR. Two radiologists evaluated 2.5- 0.625-mm-slice-thick axial images from ASIR MBIR, recorded pattern each nodule candidate, assigned confidence score. A reference standard established by consensus panel 2 different radiologists, who identified 84 noncalcified nodules diameters 4 mm or greater on reference-dose (ground-glass opacity, n = 18; partly solid, 11; 55). Sensitivity in detection assessed using McNemar test. Jackknife alternative free-response receiver operating characteristic (JAFROC) analysis applied assess results including scores.Compared CT, 78.1% decrease dose-length seen No significant differences observed between MBIR for overall sensitivity (P 0.48-0.69) JAFROC 0.57). Likewise, no ground-glass solid 0.08-0.65) 0.21-0.90).Model-based nearly an 80% radiation level level, without affecting detectability.