Evaluation of 18F-FDG PET and MRI in differentiating benign and malignant peripheral nerve sheath tumors

To compare 18F-FDG PET/CT and MRI for differentiating benign and malignant peripheral nerve sheath tumors (BPNSTs and MPNSTs) and correlate imaging characteristics with histopathology. Patients with pathologically proven PNSTs undergoing 18F-FDG PET/CT were retrospectively reviewed. PET/CTs and, if available, MRIs were analyzed, noting multiple imaging characteristics and likely pathology (benign or malignant). Thirty-eight patients with 23 BPNSTs and 20 MPNSTs were analyzed. MPNSTs had higher SUVmax (10.1 +/- 1.0, 4.2 +/- 0.4, p < 0.0001), metabolic tumor volume (146.5 +/- 39.4, 21.7 +/- 6.6 cm(3), p = 0.01), total lesion glycolysis (640.7 +/- 177.5, 89.9 +/- 23.2 cm(3)*g/ml, p = 0.01), and SUVmax/LiverSUVmean (5.3 +/- 0.5, 2.0 +/- 0.2, p < 0.0001). All lesions with SUVmax < 4.3 were benign. All lesions with SUVmax > 8.1 were malignant. SUVmax cutoff of 6.1 yielded 90.0 % sensitivity and 78.3 % specificity for MPNSTs. SUVmax/LiverSUVmean cutoff of 3.0 yielded 90.0 % sensitivity and 82.6 % specificity. MPNSTs more commonly had heterogeneous FDG activity (p < 0.0001), perilesional edema (p = 0.004), cystic degeneration/necrosis (p = 0.015), and irregular margins (p = 0.004). There was no difference in lesion size, MRI signal characteristics, or enhancement. Expertly interpreted MRI had 62.5-81.3 % sensitivity and 94.1-100.0 % specificity while PET had 90.0-100.0 % sensitivity and 52.2-82.6 % specificity for diagnosing MPNSTs. FDG PET and MRI play a complementary role in PNST evaluation. Multiple metabolic parameters and MRI imaging characteristics are useful in differentiating BPNSTs from MPNSTs. This underscores the potential critical role of PET/MRI in these patients.