Reduction of the Influence of Intravenous Contrast in PET/CT by Using a Threshold Conversion Method

In PET/CT an attenuation map for PET is derived from the CT image. However, when a full diagnostic CT with intravenous contrast is performed, the PET attenuation values are overestimated. This results in increased SUV values when compared with contrast-free studies. Although several groups have reported that the impact of contrast on the SUV values is small, the effect may not be negligible in treatment follow-up. Conventionally the hybrid scaling method is used for the conversion of CT attenuation values to PET attenuation. Here, we propose a new conversion method in an attempt to minimize the influence of the intravenous contrast on the SUV values. With the new method, CT attenuation values are converted from [0, 0.16] to [0, 0.091] (units in cm) and all the attenuation values greater than the threshold 0.16 are converted to 0.091. The threshold value of 0.16 was derived using 9 patient studies involving 14 ROIs and was further evaluated using 5 patient studies involving 13 ROIs. For the first 9 patient studies, the mean influence of the intravenous contrast on the SUV values is 5.53% (range [1.59-11.63]) with the conventional method and 2.04% (range [0.25-5.73]) with the new threshold method. Very similar results were obtained in the second group: 5.90% (range [1.29-9.59]) with the conventional method and 1.91% (range [0.04-4.59]) with the new threshold method.