SU-E-I-84: A Novel Approach for the Attenuation Correction of PET Data in PET/MR Systems.

Abstract

To evaluate the use of non-attenuated PET images (PET-NAC) as a means for the attenuation correction (AC) of PET images in PET/MR systems. A 3 step iterative segmentation process is proposed. The first step is used to segment the body contour from the NAC PET images using an active contour algorithm (Kass et al., Int J Comput Vision, 321-331 (1988)). The second step was to segment the lung region from the resultant image using an optimal thresholding approach (Xu et al., IEEE T Nucl Sci, 43, 331-336 (1996)). The purpose of the third step was to delineate parts of the heart and liver from the lung contour using a region growing approach since these parts were unavoidably included in the lung contour of the second step. Finally the attenuation coefficients of the bed were included based on CT images to eliminate the impact of the couch on the accuracy of AC. The final attenuation map was then used to AC the raw PET data and Result in a final PET image (PET-IAC). To assess the proposed segmentation approach, a phantom and six patients were scanned on a GE Discovery-RX PET/CT scanner. PET-IAC was then generated from PET- NAC using the proposed approach and compared to those of CT-AC PET (PET-CTAC). Visual inspection and SUV measurements between PET-IAC and the PET-CTAC for phantom and patient studies were performed to assess the accuracy of image quantification. Visual inspection showed a small difference between the PET-IAC and PET-CTAC. PET-IAC tumor SUVs were on average equal to 103±9% compared to the SUVs from the PET-CTAC in the phantom study, and 110±7% in the patient studies. Preliminary results suggest that PET-NAC for the AC of PET images is feasible in the clinic. Such an approach can potentially be an alternative method of MR-based AC in PET/MR imaging.