Ultra low-dose chest CT using filtered back projection: comparison of 80-, 100- and 120 kVp protocols in a prospective randomized study.

Purpose: To assess lesion detection and diagnostic image quality of filtered back projection (FBP) reconstruction technique in ultra low-dose chest CT examinations.

Methods And Materials: In this IRB-approved ongoing prospective clinical study, 116 CT-image-series at four different radiation-doses were performed for 29 patients (age, 57-87 years; F:M - 15:12; BMI 16-32 kg/m(2)). All patients provided written-informed-consent for the acquisitions of additional ultra low-dose (ULD) series on a 256-slice MDCT (iCT, Philips Healthcare). In-addition to their clinical standard-dose chest CT (SD, 120 kV mean CTDIvol, 6 ± 1 mGy), ULD-CT was subsequently performed at three-dose-levels (0.9 mGy [120 kV]; 0.5 mGy [100 kV] and 0.2 mGy [80 kV]). Images were reconstructed with FBP (2.5mm 1.25 mm) resulting into four-stacks: SD-FBP (reference-standard), FBP0.9, FBP0.5, and FBP0.2. Four thoracic-radiologists from two-teaching-hospitals independently-evaluated data for lesion-detection and visibility-of-small-structures. Friedman's-non-parametric-test with post hoc Dunn's-test was used for data-analysis.

Results: Interobserver-agreement was substantial between radiologists (k=0.6-0.8). With pooled analysis, 146-pulmonary (27-groundglass-opacities, 64-solid-lung-nodules, 7-consolidations, 27-emphysema) and 347-mediastinal/soft tissue lesions (87-mediastinal, 46-hilar, 62-axillary-lymph-nodes, and 11-mediastinal-masses) were evaluated. Compared to the SD-FBP, 100% pulmonary-lesions were seen with FBP0.9, up to 81% with FBP0.5 (missed: 4), and up to 30% with FBP0.2 images (missed:16). Compared to SD-FBP, all enlarged mediastinal-lymph-nodes were seen with FBP0.9 images. All mediastinal-masses (>2 cm, 11/11) were seen equivalent to SD-FBP images at 0.9 mGy. Across all sizes of patients, FBP0.9 images had optimal visualization for lung findings. They were optimal for mediastinal soft tissues for only non-obese patients.

Conclusion: Filtered-back-projection technique allows optimal lesion detection and acceptable image quality for chest-CT examinations at CDTIvol of 0.9 mGy for lung and mediastinal findings in selected sizes of patients.