18F-FDG PET/CT characterization of adrenal lesions

Abstract

647

Objectives: Adrenal lesions are not uncommon. However, not all adrenal lesions are metastatic and up to 9% of the general population harbors benign adenoma. Reported accuracies of different imaging techniques in adrenal lesions characterizations are quite variable. Although extensive research has been performed in regard to the diagnostic performance of PET alone and CT alone for the detection of adrenal lesion, the optimal imaging characterization strategy has not been defined. The aim of this study was to define the best cut off combination of mean standard uptake value (SUVmean) and mean Hounsfield units (HUmean) that best characterize adrenal lesions.

Methods: 20 volunteers with normal adrenal glands (mean age 42, 16 female, 4 male) and 47 cancer patients (mean age 61, 27 female, 30 male) were studied. Of the 47 patients, 14 had metastatic adrenal lesions and 33 had either normal and/or benign adrenal glands, all proven by pathology, clinical or imaging follow-up. F-18 FDG PET/CT was performed on a Gemini, Philips Medical Systems applying a non-contrast, low-energy CT. SUVmean and HUmean were obtained after drawing a representative free-hand region of interest of the adrenal glands or lesions. Receiver operating characteristic curve analysis was used to determine cut-off points of SUVmean and HUmean for the detection of malignancy.

Results: Summaries of statistical results are included in Table 1 and Figure 1. The SUVmean provides a higher area under the curve (AUC: 0.989) compared to HUmean (AUC: 0.898) suggesting better sensitivity and specificity for the detection of malignancy.

Conclusions: We conclude that using SUVmean ≥ 2.2 provides higher sensitivity, specificity, positive and negative predictive value compared to HUmean. Moreover, using a cut off combination of SUVmean ≥ 2.2 and a HUmean ≥20 provides the highest specificity, positive predictive values and accuracy in adrenal lesion characterization. Prospective validation of this SUV-HU combination is in progress.

Table 1: Summary of statistical results