Background: The appearance of malignant peripheral nerve sheath tumours (MPNST) marks a critical stage in the course of neurofibromatosis type 1 (NF1). Since the diagnostic criteria are fairly non-specific, histological examination alone can confirm malignancy. We assessed the value of 18 FDG positron emission tomoscintigraphy (PET scan) in screening for such malignant tumours.
Patients and methods: Between October 2000 and August 2006, all of our patients with NF1 and suspected MPNST underwent PET scan. Inclusion criteria consisted of clinical signs (increased tumour size or induration, pain) and/or laboratory values. Analysis of PET scan images, based upon determination of tumour/liver binding ratio with a cut-off point of 1.5 times hepatic binding, was used to classify lesions as non-suspect or pathological. In the case of suspect lesions, histological analysis was performed. For non-suspect lesions, patients either underwent monitoring or excision of the lesion where necessary and technically feasible.
Results: Thirty-eight patients with 49 tumours were included in the study. In 8 patients, PET scan showed suspect lesions (12 tumours), and histological analysis of these tumours revealed 6 MPNST. In 30 patients (37 tumours) PET scan showed non-suspect binding, and no malignant tumours were demonstrated either on histological examination or after mean follow-up of 33.5 months. PET scan thus demonstrated a sensitivity and negative predictive value of 100%, specificity of 86%, and a positive predictive value of 50%. The ratios of positive and negative probability were respectively 7.14 and 0.
Discussion: Our study, to our knowledge the most extensive yet performed, demonstrates the value of PET scan in detecting MPNST, particularly based on its 100% negative predictive value. To date, other than biopsy, no examinations allow diagnosis with any certainty. The literature reports 2 studies analysing the value of PET scan. The first involved 18 NF1 patients with 23 plexiform neurofibromas: of 7 tumours with hyperbinding, 5 were MPNST. The second study concerned 5 NF1 patients with a total of 15 tumours: 7 tumours showed hyperbinding, of which 6 were MPNST. Using the same evaluation criteria, our study yielded comparable results. The negative predictive value of 100% provides a strong argument in favour of a benign tumour.
Conclusion: Our study confirms the value of PET scan in the detection of NF1 even though false positives require medical-surgical confirmation before any potentially detrimental therapeutic decisions may be made.