18F-Choline PET/CT, MRI, and Software-Based Image Fusion Analysis in Patients With Primary Hyperparathyroidism

Abstract

Purpose The aim of this study was to evaluate the diagnostic performance of 18F-choline PET and MRI in patients with primary hyperparathyroidism. Furthermore, the additional value of software-based PET/MRI scan fusion was analyzed. Patients and Methods This retrospective study includes 42 patients (38 women) with an age between 32.5 and 79.1 years. PET/CT scans were performed on a dedicated system after injection of 250 to 350 MBq 18F-choline. For the MRI examination, T1-weighted images of the cervical region were used. The image fusion was made by anatomical coregistration using an automated algorithm based on mutual information. Results A total of 46 lesions were discovered and histologically confirmed in 42 patients. Histopathological examination revealed 38 adenomas and 8 hyperplasias. This means that, in 4 of these 42 patients, 2 lesions per patient were discovered. PET/CT also detected 46 abnormal findings, but only 43 were correctly recognized, whereas the other 3 were false-positive (FP). Six lesions could not be detected correctly: 3 were FP and 3 false-negative, which resulted in a sensitivity of 93.5% and a specificity of 97.5%. The site-specific evaluation showed 18 true-positive enlarged parathyroid glands with MRI, but also produced 13 FP findings and failed to detect 28 lesions; the sensitivity and specificity are thus 39.1% and 89.3%, respectively. The difference in detection rate between 18F-choline PET/CT and MRI was statistically significant (P < 0.001). Conclusions 18F-choline PET/CT is clearly superior to MRI for localization diagnostics in primary hyperparathyroidism. Image fusion of both modalities can be helpful for more precise anatomical assignment.