Standardised uptake values from PET/CT images: comparison with conventional attenuation-corrected PET

Abstract

In PET/CT, CT-derived attenuation factors may influence standardised uptake values (SUVs) in tumour lesions and organs when compared with stand-alone PET. Therefore, we compared PET/CT-derived SUVs intra-individually in various organs and tumour lesions with stand-alone PET-derived SUVs. Thirty-five patients with known or suspected cancer were prospectively included. Sixteen patients underwent FDG PET using an ECAT HR+scanner, and subsequently a second scan using a Biograph Sensation 16PET/CT scanner. Nineteen patients were scanned in the reverse order. All images were reconstructed with an iterative algorithm (OSEM). Suspected lesions were grouped as paradiaphragmatic versus distant from the diaphragm. Mean and maximum SUVs were also calculated for brain, lung, liver, spleen and vertebral bone. The attenuation coefficients (mu values) used for correction of emission data (bone, soft tissue, lung) in the two data sets were determined. A body phantom containing six hot spheres and one cold cylinder was measured using the same protocol as in patients. Forty-six lesions were identified. There was a significant correlation of maximum and mean SUVs derived from PET and PET/CT for 14 paradiaphragmatic lesions (r=0.97 respectively; p<0.001 respectively) and for 32 lesions located distant from the diaphragm (r=0.87 and r=0.89 respectively; p<0.001 respectively). No significant differences were observed in the SUVs calculated with PET and PET/CT in the lesions or in the organs. In the phantom, radioactivity concentration in spheres calculated from PET and from PET/CT correlated significantly (r=0.99; p<0.001). SUVs of cancer lesions and normal organs were comparable between PET and PET/CT, supporting the usefulness of PET/CT-derived SUVs for quantification of tumour metabolism.