• search hit 7 of 8
Back to Result List

Computed tomography lung-cancer screening: does iterative reconstruction matter?

  • Background: By reviewing image quality and diagnostic perception, the suitability of a statistical model-based iterative reconstruction algorithm in conjunction with low-dose computed tomography for lung cancer screening is investigated. Methods: Artificial lung nodules shaped as spheres and spiculated spheres made from material with calibrated Hounsfield units were attached on marked positions in the lung structure of anthropomorphic phantoms. The phantoms were scanned using standard high contrast, and two low-dose computed tomography protocols: low-dose and ultra-low-dose. For the reconstruction, the filtered back projection and the iterative reconstruction algorithm ADMIRE at different strength levels (S1–S5) and the kernels Bl57, Br32, Br69 were used. Expert radiologists assessed image quality by performing 4-field-ranking tests and reading all image series to examine the aptitude for the detectability of lung nodules. Signal-to-noise ratio was investigated as objective image quality parameter. Results: In ranking tests for lung foci detection expert radiologists prefer medium to high iterative reconstruction strength levels. For the standard clinical kernel Bl57 and varying phantom diameter, a noticeable preference for S4 was detected. Experienced radiologists graded filtered back projection reconstructed images with the highest perceptibility. Less experienced readers assessed filtered back projection and iterative reconstruction equally with the highest grades for the Bl57 kernel. Independently of the dose protocol, the signal-to-noise ratio increases with the iterative reconstruction strength level, specifically for Br69 and Bl57. Conclusions: Subjective image perception does not significantly correlate with the experience of the radiologist, which presumably mirrors reader’s training and accustomed reading adjustments. Regarding signal-to-noise ratio, iterative reconstruction outperforms filtered back projection for spheres and spiculated spheres. Iterative reconstruction matters. It promises to be an alternative to filtered back projection allowing for lung-cancer screening at markedly decreased radiation exposure but comparable or even improved image quality.

Export metadata

Additional Services

Share in Twitter Search Google Scholar
Metadaten
Author:Britta König, Hilmar Kühl, Waldemar Zylka
DOI:https://doi.org/10.21037/shc-21-9
ISSN:2521-3768
Parent Title (English):Shanghai Chest
Document Type:Article
Language:English
Date of Publication (online):2021/12/17
Year of first Publication:2021
Publishing Institution:Westfälische Hochschule Gelsenkirchen Bocholt Recklinghausen
Release Date:2021/12/17
Volume:5
Issue:October
First Page:Artikelnr. 35
Departments / faculties:Fachbereiche / Elektrotechnik und angewandte Naturwissenschaften
Licence (German):License LogoEs gilt das Urheberrechtsgesetz

$Rev: 13159 $