Experimental evaluation of the bilinear transformation used in the CT-based attenuation correction for small animal PET imaging

The purpose of this work is to experimentally evaluate the accuracy of the bilinear transformation method for generating attenuation maps, the accuracy of the CT-based attenuation correction (AC) using these maps for small animal PET imaging, and the reliability of this method when an iodinated contrast agent is introduced in CT scans. The attenuation maps of an attenuation calibration phantom (ACP) with known linear attenuation coefficients (LAC) were generated from both Ge-68 transmission scan and bilinear transformation of CT scan. These maps were then used in AC for quantitative analysis of a hot small sphere imbedded in the ACP phantom. We measured LAC of the mouse blood mixed with an iodinated agent (Fenestra VC) using both Ge-68 transmission scan and bilinear approach, and investigated the effect of CT contrast agent on CT images, attenuation maps and attenuation corrected PET images of the mice, respectively. The phantom experiments showed visually well-matched attenuation maps and within 5% bias agreement on LAC of artificial tissues. CT-based and Ge-transmission-based AC restored the quantitative PET results for the hot small sphere in ACP phantom from -27% error (without AC) to 0.9% and -3.4%, respectively. The animal experiments showed that CT-based AC restored respectively 15-25% and 40- 60% PET image intensities for mouse and rat tissues. With the increasing of the ratio of Fenestra to blood, the difference between the CT-based and Ge-transmission-based LAC of the mixture was increased linearly up to 7% at the maximum ratio, which corresponds to the routine limitation on intravenous injection dose for a mouse. The administration of 0.012 ml/gm Fenestra VC increased 2.1plusmn1.2%, 0.6plusmn0.4%, 0.2plusmn0.2%, 0.4plusmn0.3%, and 0.1plusmn0.2% mean image intensities respectively for the heart chamber, liver, brain, cortical muscle and bladder on PET images with CT-based AC. We conclude that the water-bone- based bilinear approach is adequate for generati- ng attenuation maps for CT-based AC on small animal PET imaging. The administration of Fenestra VC under the routine dosage (<0.016 ml/gm) results in visible (<7%) deviation on LAC of blood, but negligible (<3% and <1%) over-corrections on mouse heart and other tissues in PET imaging using bilinear-approach- based CT-AC.