Comparative Effectiveness of 18F-FDG PET/CT Versus Whole-Body MRI for Detection of Malignant Peripheral Nerve Sheath Tumors in Neurofibromatosis Type 1

Authors:
  • KTKatharina Tornquist
    Katharina Tornquist
  • SMSilvia Münster
    Silvia Münster
  • IAIvayla Apostolova
    Ivayla Apostolova
  • CHChristian Hagel
    Christian Hagel
  • RFReinhard E. Friedrich
    Reinhard E. Friedrich
  • UWUlrike Wedegärtner
    Ulrike Wedegärtner
  • VMVictor F. Mautner
    Victor F. Mautner

Abstract
The aim of this study was to compare the diagnostic performance of ¹⁸F-fluorodeoxyglucose (FDG) PET/CT and whole-body MRI for the detection of malignant peripheral nerve sheath tumors (MPNSTs) in patients with neurofibromatosis type 1, and to evaluate a panel of imaging-based criteria serving that purpose. Thirty-one patients were examined by whole-body MRI and ¹⁸F-FDG PET/CT. A panel of imaging-based criteria including tumor region, size, shape, margin definition, contrast enhancement, heterogeneity before and after contrast, intratumoral lobulation, target sign, and mean and maximum standardized uptake values (SUVs) were evaluated. A SUVmax cut-off value of 3.5 was used for lesion analysis. Histopathologic evaluation and/or clinical follow-up served as the reference standard. ¹⁸F-FDG PET/CT had a sensitivity of 100%, whereas MRI had a sensitivity of 66.7%. On PET/CT, tumor size (P<0.005), SUVmax (P<0.0001), SUVmean (P<0.0001), and tracer uptake heterogeneity (P=0.002) were significantly associated with MPNSTs. On MRI, intratumoral lobulation (P<0.02), ill-defined margins (P=0.007), and irregular enhancement on T1-weighted imaging (P<0.001) were significantly associated with MPNSTs. Both PET/CT and whole-body MRI may distinguish benign and malignant PNSTs, but PET/CT has higher sensitivity for that purpose. Imaging-based criteria for identification of MPNSTs on both modalities were identified. False-positive results, requiring biopsy or clinical follow-up, may be reduced by using a combination of MRI and PET derived markers, but only at the price of reduced sensitivity.

This publication has 21 references indexed in Scilit: