Alain Vlassenbroek PhD, Presenter: Employee, Koninklijke Philips NV

Domitille Millon MD, Abstract Co-Author: Nothing to Disclose

Emmanuel Etienne Coche MD, Abstract Co-Author: Nothing to Disclose

To compare the visibility of structures of various sizes and contrast in MDCT images reconstructed with an iterative reconstruction algorithm and with a filtered back projection algorithm in standard and high resolution at various dose levels

The experimental study was performed on a 256-slice MDCT (Philips Healthcare, Cleveland, OH). Modulation transfer functions (MTF) were measured from the edges of the 4 sensitometry samples of the CTP401 module of a catphan phantom (The phantom Laboratory, Salem, NY). The CT images of the various contrast inserts were scanned at decreasing doses (48.8 mGy down to 0.7 mGy) and reconstructed with standard filtered back projection (FBP) and iterative reconstruction algorithm (IMR, Philips Healthcare). The edge of each circular contrast object was analyzed to determine the edge spread function, which was differentiated and Fourier transformed to obtain the object-specific MTF. IMR results were compared to the FBP results at the various dose and contrast levels to analyze the influence of those parameters on the spatial resolution. In addition, chest cadaver images scanned and reconstructed using the same technical parameters for decreasing dose levels and were analyzed to show the differences in the visibility of anatomical structures between the 2 reconstruction techniques and to compare the results obtained with the phantom scans.

With FBP, the MTF measured from all contrast inserts was constant throughout all measurements. With IMR, the MTF measured at the highest dose was similar to that of FBP. However, with IMR, the MTF was increasingly lowered for inserts of decreasing contrast at decreasing dose (LDPE insert: MTF50% = 6.4, 5.95, 4.3, 2.9 lp/cm at 48.8, 6.1, 3.0 and 0.7 mGy respectively). Cadaver images reconstructed with IMR showed similarly that the visibility and delineation of anatomical structures could be deteriorated with decreasing doses, and that this happened primarily with the smallest and lowest contrast structures.

Results demonstrated that although iterative reconstructions provide a superior noise performance to FBP, spatial resolution of iterative reconstructions is dose and contrast dependent.

Clinical scan protocols which had been optimized based on the CT noise properties of FBP will need to be re-evaluated with iterative reconstructions using new image quality metrics based on this new paradigm shift.

Vlassenbroek, A, Millon, D, Coche, E, Dose and Contrast Dependence of Spatial Resolution with Iterative Reconstructions of Multidetector Computed Tomography Images: A Phantom and Cadaver Study.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14007078.html