18F-FDG PET/CT qualitative and quantitative evaluation in neurofibromatosis type 1 patients for detection of malignant transformation: comparison of early to delayed imaging with and without liver activity normalization.

Abstract

(18)F-FDG PET/CT has shown increased accuracy, compared with morphologic imaging, in differentiating malignant peripheral nerve sheath tumors (MPNSTs) from benign neurofibromas (BNFs) in patients with neurofibromatosis type 1 (NF1). Delayed (18)F-FDG PET imaging typically enhances malignant tumor to background. Our goal was to compare the effectiveness of early (1-h) and delayed (4-h) (18)F-FDG PET/CT imaging in differentiating MPNSTs from BNFs in patients with NF1, with and without liver activity normalization. NF1 patients presenting new symptoms or enlarging lesions were clinically evaluated with early and delayed (18)F-FDG PET/CT imaging. SULmax (maximum standardized uptake value derived for lean body) and SULmax/liver (lesion uptake adjusted to mean liver activity) were obtained for all sites identified with abnormal metabolic activity. Qualitative and quantitative evaluations, including receiver-operating-characteristic (ROC) comparison of early and delayed imaging sessions, were performed. Histopathology and clinical follow-up (1-9 y) were considered as a gold standard. Forty-one NF1 patients with early and delayed (18)F-FDG PET/CT scans were identified, and 93 lesions were retrospectively analyzed, representing 24 MPNSTs (all histologically confirmed) and 69 BNFs (26 histologically confirmed). Qualitative evaluation on early imaging showed sensitivity, specificity, positive predictive value, and negative predictive value for separating MPNSTs from BNFs of 91%, 84%, 67%, and 96% versus 91%, 81%, 63%, and 96%, respectively, on 4-h delayed imaging. The mean SULmax was significantly higher for MPNSTs than BNFs on both early scans (6.5 vs. 2.0, P < 0.01) and delayed imaging (8.3 vs. 2.3, P < 0.02). However, SULmax overlap between benign and malignant lesions persisted even after normalization to mean liver activity. ROC-derived best SULmax cutoffs were 3.2 on early (area under the curve, 0.973) and 4.1 on delayed scans (area under the curve, 0.978). ROC analysis for SULmax/liver improved test specificity (94% vs. 87%, P < 0.05) on early and (93% vs. 88%, P < 0.05) on delayed imaging. Qualitative interpretation of (18)F-FDG PET/CT discriminates MPNSTs from BNFs in NF1 patients with similar accuracy on both early and delayed imaging. Quantitative data showed better sensitivity on delayed acquisition and best test specificity with lesion SULmax normalization to liver activity, more so than with delayed imaging at 4 h.