RSNA 2011 

Abstract Archives of the RSNA, 2011


SST16-08

Impact of 4th Generation Iterative Reconstruction Technique on Image Quality in Low Dose Computed Tomography of the Lungs

Scientific Formal (Paper) Presentations

Presented on December 2, 2011
Presented as part of SST16: Physics (CT Dose and Reconstruction)

Participants

Jan-Hendrik Buhk MD, Abstract Co-Author: Nothing to Disclose
Azien Laqmani, Abstract Co-Author: Nothing to Disclose
Hans-Christoph von Schultzendorff, Abstract Co-Author: Nothing to Disclose
Diego Hammerle, Abstract Co-Author: Nothing to Disclose
Frank Oliver Henes MD, Abstract Co-Author: Nothing to Disclose
Michael Groth MD, Abstract Co-Author: Nothing to Disclose
Hans Dieter Nagel, Abstract Co-Author: Consultant, Koninklijke Philips Electronics NV
Gerhard B. Adam MD, Abstract Co-Author: Nothing to Disclose
Marc Regier, Presenter: Nothing to Disclose

PURPOSE

To intraindividually compare and evaluate the influence of the 4th generation iterative reconstruction (IR) technique iDose™ on low-dose MDCT of the lungs with regard to anatomical delineation and artifact reduction.

METHOD AND MATERIALS

In 29 consecutive patients suspected of suffering from infectious pneumonia the raw data of unenhanced MDCT scans (120 kV, standard resolution mode, mean CTDIvol: 2.6 mGy, mean DLP: 125 mGy x cm; 17 male, 12 female; mean age 47y) were reconstructed using a prototypic reconstruction processor featuring iDose4™ software (Philips, Best, the Netherlands). iDose™ level settings were varied between 0, 2, 4, 6 and 7. Additionally, for each setting three different thorax-specific filter kernels were applied (L, lung; B, standard; C, sharp). Images were reconstructed with a slice thickness of 1 mm (L) or 5 mm (B, C), respectively. Three radiologists blinded to the reconstruction technique independently performed image analysis. A 4-point grading scale was applied regarding the depiction of anatomical details of the lungs and mediastinal structures and the degree of artifacts observed (1, worst; 4, excellent). The signal-to-noise ratio (SNR) was assessed for each data set. Statistical evaluation included weighted kappa-analysis and Wilcoxon-test.

RESULTS

The highest image quality regarding the lung was assessed applying an iDose™ level of 4 (kernel L). Regarding mediastinal soft tissue iDose™ level of 4 (kernel C) scored best. Even at the lower iDose™ level of 2 the image ratings scored superior to standard filtered back-projection (= iDose™ level 0). Inter-rater reliability was at least substantial in all cases (κ > 0.6). With increasing iDose™ levels, reduction of streak artifacts was visible and successfully quantified by a decrease of the standard deviation at constant mean attenuation values.

CONCLUSION

Iterative reconstruction seems to be a powerful tool for improvement of image quality in low-dose MDCT of the lung and allows better depiction of the mediastinal structures as well. Even at relatively low iDose™ levels higher image quality and substantial reduction of streak artifacts compared to the filtered back-projection can be achieved.

CLINICAL RELEVANCE/APPLICATION

The application of a 4th generation iterative reconstruction technique has the potential to remarkably improve image quality in low-dose MDCT of the lung and mediastinum.

Cite This Abstract

Buhk, J, Laqmani, A, von Schultzendorff, H, Hammerle, D, Henes, F, Groth, M, Nagel, H, Adam, G, Regier, M, Impact of 4th Generation Iterative Reconstruction Technique on Image Quality in Low Dose Computed Tomography of the Lungs.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11016753.html