No diagnostic impact of routinely use of respiratory gating to characterize lung nodules with 18F-FDG PET/CT

Abstract

PurposeTo evaluate diagnostic impact of routinely use of respiratory gated (RG) 18FDG PET/CT to distinguish benign and malignant lung nodules. MethodsWe analyzed retrospectively data of 76 patients referred for metabolic characterization of lung nodules who underwent whole body 18FDG PET/CT scan followed by RG PET/CT (deep-inspiration breath-hold, DIBH). RG was prospectively planned without knowledge of WB PET/CT results. Independent reading of PET/CT scans with or without respiratory gating was separately conducted by two nuclear medicine physicians. Uptake of lung nodules was evaluated visually (positive if nodule's uptake superior to pulmonary background uptake), and quantitatively (SUVmax, SUVmean, metabolic volume, tumor to background ratio). Reference standard was obtained for all patients by histology (n=37) or clinico-radiological follow-up (n=39). Sub-groups were also evaluated (nodule<15mm, lower lobe nodules). ResultsNodules were classified positives without RG scans in 43/76 (59.7%) patients and with RG scans in 41/75 (54.6%) patients. Results were discordant for 5/75 (6.7%) patients. According to the reference standard, 35/39 cancers were correctly identified by PET/CT without RG, and 34/38 by RG PET/CT. Without RG PET, sensitivity, specificity and accuracy were 89.7%, 78.4% and 84.2% respectively. With RG PET, sensitivity, specificity and accuracy were 89.5%, 81.1% and 85.3% respectively, not statistically different. Concerning quantitative analysis, results were not statistically better with RG than without RG and were not better than visual analysis. Sub-groups analysis showed no added value of RG scans in specific groups (lower lobes and smaller nodules). ConclusionRoutinely use of RG did not help in the diagnosis of neoplastic lung nodules. Others evaluations are needed to assess the contribution of RG for others selective indications (therapeutic evaluation, radiotherapy planning, characterization of liver lesions).