P.D. Bamidis and N. Pallikarakis (Eds.): MEDICON 2010, IFMBE Proceedings 29, pp. 220–223, 2010. www.springerlink.com A Novel Approach for Implementation of Dual Energy Mapping Technique in CT-Based Attenuation Correction Using Single kV P Imaging: A Feasibility Study B. Teimourian 1,2 , M.R. Ay 2,3,4 , H. Ghadiri 2,5 , M. Shamsaei Zafarghandi 1 , and H. Zaidi 6,7 1 Faculty of Physics and Nuclear Engineering, Amir Kabir University of Technology (Tehran Polytechnic), Tehran, Iran 2 Research Center for Science and Technology in Medicine, Tehran University of Medical Sciences, Tehran, Iran 3 Department of Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences, Tehran, Iran 4 Research Institute for Nuclear Medicine, Tehran University of Medical Sciences, Tehran, Iran 5 Department of Medical Physics, Iran University of Medical Sciences, Tehran, Iran 6 Geneva University Hospital, Division of Nuclear Medicine, Geneva, Switzerland 7 Geneva University, Geneva Neuroscience Center, Geneva, Switzerland Abstract— In the CT-based attenuation correction methods, dual-energy technique (DECT) is the most accurate approach, which has been limited due to the increasing patient dose. In this feasibility study, we have introduced a new method that can implement dual-energy technique with only a single energy CT scan. In this method, with having the CT image in one energy, we generate the CT image at the second energy (from now we call it virtual dual-energy technique). The attenuation map at 511 keV was generated using bilinear (the most commonly used method in commercially available PET/CT scanners), dual-energy and virtual dual-energy technique in phantom and patients data. In the phantom study, the created attenuation map using mentioned methods are compared to the theoretical values calculated using XCOM cross section library. In the patient study, the generated attenuation map using dual-energy method is considered as gold standard. The results in the phantom data show 10.1 %, 4.2 % and 4.3 % errors for bilinear, dual-energy and virtual dual- energy techniques respectively. Also, the results in the patient data show the virtual dual-energy has better agreement with the dual-energy method rather than the bilinear method especially in the bone tissue (1.5 % and 8.9 % respectively). Keywords— PET/CT, DECT, attenuation correction, at- tenuation map, energy-mapping. I. INTRODUCTION Hybrid positron emission tomography/computed tomogra- phy (PET/CT) units have been designed and been commer- cially available since 2000 [1]. The additional morphological information provided by PET/CT scanners in contrast to stand alone PET scanners can be of additional diagnostic value for the physicians. Another benefit of PET/CT systems is the faster examination time, since the attenuation map for PET data correction is obtained from the CT scan and not from the much longer transmission scan [2]. Although fast and precise CT-based attenuation correc- tion (CTAC) method yields a noise free attenuation map in comparison with transmission scan, but CT images pro- vide linear attenuation coefficients (LAC) of the tissues at effective CT energies (~ 60-80 keV) rather than 511 keV which is the energy of PET imaging, so it is necessary to convert the LAC at CT energies to those corresponding to 511 keV [3]. Several energy mapping strategies including scaling [4], segmentation [4], hybrid (Segmentation and scaling) [4], bilinear [5] and dual-energy (DECT) [6] have been proposed that convert the LACs of CT images to the LACs of 511 keV. It should be noted that most commer- cially available PET/CT scanners use the bilinear method. The best of these mentioned methods is dual energy technique [7]. Some of the drawbacks of this method that render it impractical for commercial PET/CT scanners are the additional dose to the patient, resulting from two CT scans at two different kV P s, increasing the scanning time as well as cost. In this feasibility study we have introduced a new method that can implement dual-energy technique with only a single energy (kV P ) CT imaging. In this method, with having the CT image in one energy, we generate the CT image at the second energy. II. MATERIALS AND METHODS A. kV P Conversion Curves The Alderson RANDO (Radiology Support Devices Company, USA) phantom [8] was scanned by a LightSpeed VCT scanner (GE Healthcare Milwaukee, USA) with four different tube voltages (80, 100, 120, and 140 kV P ) and tube current of 300 mA. The analysis on the acquired images was done by AMIDE [9] image viewer. More than 400 different ROIs were selected in each image and the mean CT numbers for each ROI at one kV P was plotted versus the same values at another kV P . Finally the best curve was fitted for each plot to obtain kV P Conversion Curves (which scale CT numbers at different tube voltages to each other), in three regions includ- ing lung tissue (HU<-100), soft tissue (-100<HU<200) and bone tissue (HU>200). This classification improves the preci- sion of the resulted kV P conversion curves.