Daniele Marin MD, Presenter: Fellowship funded, Bracco Group

Rendon C. Nelson MD, Abstract Co-Author: Consultant, General Electric Company Research support, Bracco Group Research support, Becton, Dickinson and Company Speakers Bureau, Siemens AG

Daniela Barbara Husarik MD, Abstract Co-Author: Nothing to Disclose

Samuel Richard PhD, BSC, Abstract Co-Author: Nothing to Disclose

James G. Colsher PhD, Abstract Co-Author: Employee, General Electric Company

Ehsan Samei PhD, Abstract Co-Author: Advisory Board, Ion Beam Applications, SA Consultant, Siemens AG

Sebastian Tobias Schindera MD, Abstract Co-Author: Research grant, Siemens AG

Terry T. Yoshizumi PhD, Abstract Co-Author: Research support, General Electric Company

Erik K. Paulson MD, Abstract Co-Author: Research Consultant, Siemens AG Stockholder, ZONARE Medical Systems, Inc

To investigate the impact of a model-based iterative reconstruction algorithm (MBiR; GE Healthcare) on the selection of tube potential (kV) to decrease radiation dose while preserving image quality for the detection of hypervascular liver tumors at multi–detector row CT (MDCT).

A customized, iodine-containing, anthropomorphic liver phantom (CIRS, Norfolk, Va) was manufactured to simulate hepatic attenuation at different tube voltages during the late arterial phase (from +70 HU at 140 kVp to +104 HU at 80 kVp). Iodinated solid spheres measuring 5, 10, and 15 mm in diameter were used to simulate slightly hypervascular liver tumors with various degrees of enhancement. The liver phantom was scanned with a 64-section MDCT scanner (Discovery CT750 HD; GE Healthcare) at 120, 100, and 80 kVp. Tube current–time product settings (mAs) were adjusted to maintain clinically adequate noise levels (noise index, 14-19; mA range, 80-650) across the entire kVp spectrum. For each kVp, mAs settings were then incrementally decreased to 75%, 50%, 25%, and 10% of their baseline value. All other imaging parameters were kept constant. Each protocol was repeated three times for consistency. CT raw data were reconstructed with both standard convolution filtered back projection (FBP) and MBiR algorithms. Image noise and lesion-to-liver contrast-to-noise ratio (CNR) for the largest (15 mm) lesions were calculated and compared using a conservative noise threshold of 20 HU. Radiation dose was compared using volume-averaged CT dose index (CTDI).

At equal kVp and mAs settings, MBiR yielded 56%-535% CNR increase and 41%-87% noise reduction compared with FBP (P <0.001 for all comparisons). At each kVp setting, MBiR enabled up to 90% (10 fold) radiation dose reduction without exceeding the 20 HU noise threshold. MBiR-reconstructed 80 kVp images yielded the highest lesion conspicuity, reflected by 568% CNR increase and 37% radiation dose reduction compared to our standard 120 kVp CT technique (280 mAs, FBP).

The use of low tube voltage (80 kVp) technique combined with MBiR algorithm may increase the conspicuity of hypervascular liver tumors while significantly reducing radiation dose.

Judicious selection of kV, along with the use of a MBiR algorithm, can dramatically decrease radiation dose and potentially improve lesion detection rates during contrast-enhanced MDCT of the liver

Marin, D, Nelson, R, Husarik, D, Richard, S, Colsher, J, Samei, E, Schindera, S, Yoshizumi, T, Paulson, E, Detection of Hypervascular Liver Tumors at MDCT: Optimization of Tube Potential for Radiation Dose Reduction and Image Quality Improvement Using a Model-based Iterative Reconstruction (MBiR) Algorithm.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9009325.html