PET-CT: accuracy of PET and CT spatial registration of lung lesions

PET-CT scanners offer the unique ability to acquire PET and CT data with rapid full body registration. The purpose of this study was to evaluate the accuracy of spatial registration between PET and CT data in patients with fluorine-18 fluoro-2-deoxy-D-glucose (FDG)-avid lung lesions. PET, CT and fused PET-CT images from 244 consecutive clinical patients undergoing whole-body FDG PET-CT imaging (GE Discovery LS, CT attenuation correction, OSEM reconstruction) were evaluated. Inclusion criteria for this analysis were lung lesions clearly defined on both PET and CT, lesion diameter less than 5 cm and clear borders. Patients were allowed to breathe freely during both PET and CT image acquisitions. The spatial coordinates of the visually estimated centers of the lesion were determined independently for PET and CT images and compared. Thirty-six patients (26 females, 10 males) with a total of 48 lesions were included (19 lung cancer patients with 26 lesions, 17 patients with 22 lung metastases). The average lung lesion diameter was 15.6 +/- 9 mm. The mean-distance between the center of lesions independently determined for both PET and CT was 7.55 +/- 4.73 mm. Misregistration tended to be more pronounced in the lower lungs (10.2 +/- 6.55 mm) than in the upper lungs (6.67 +/- 4.28 mm) (P=0.063). Misregistration also tended to be slightly more pronounced in the left lung (8.33 +/- 5.05 mm) than in the right lung (6.25 +/- 3.92 mm) (P=0.059). In conclusion, with a dedicated PET-CT scanner and this clinically practical imaging algorithm, registration is usually accurate, but spatial misregistration of primary lung lesions does occur.