Aim: Bronchioloalveolar (BAC) cancer is a source of false-negative F-fluorodeoxyglucose (FDG) PET/computed tomography (CT) results. A few studies reported better diagnostic performances with PET tracers of lipid metabolism, C-choline, or C-acetate, for the detection of well-differentiated adenocarcinoma or BAC. F-fluorocholine (FCH) is a lipid analogue for PET imaging, with advantages in terms of logistics and image resolution. We carried out this prospective pilot study to evaluate whether FCH PET/CT could detect lung cancer with a BAC component and could be more sensitive than FDG in this aim.
Methods: Fifteen patients with a lung nodule or lesion suspected for BAC on CT and/or with a history of BAC had PET/CT 60-90 min after 5 MBq FDG/kg body mass and, on a separate day, 10-20 min after 4 MBq FCH/kg body mass. The standard of truth was histology and a 6-month follow-up.
Results: Nine patients (12 lesions) presented BAC or adenocarcinoma with BAC features, two patients presented adenocarcinoma without BAC features (five lesions) and four patients presented benign lesions (15 non-malignant sites). For both FCH and FDG, patient-based sensitivity was 78% for detecting cancer with a BAC component and 82% for detecting malignancy. Site-based sensitivity for detecting malignancy was 76 and 75% for detecting cancer with BAC features, for both radiopharmaceuticals. Specificity was similar for FCH and FDG (site-based 93 vs. 81%, NS). In these early-stage cancers, only one adrenal metastasis was observed that took up FCH and FDG.
Conclusion: In this population of patients with ground-glass opacities selected on CT suggestive of BAC or with a history of BAC and a recent lung anomaly on CT, FCH detected all malignant lesions with at least a 2.0 cm short axis. However, FDG had similar performance.