173 A Clinical and Radiographic Score to Assess Malignant Potential of Peripheral Nerve Sheath Tumors.

INTRODUCTION: The differentiation of malignant peripheral nerve sheath tumors (MPNST) from benign nerve sheath tumors is critical to direct treatment. Although histopathological diagnosis remains the standard, clinical and radiographic assessments are key for lesions that present with diagnostic uncertainty. Previous studies exploring these criteria have been limited by low sensitivity and specificity. Here, we describe a scoring system based on clinical presentation, neurological examination, and MRI and apply it retrospectively to a single-surgeon series of peripheral nerve sheath tumors.

METHODS: Clinical characteristics common to the presentation of peripheral nerve sheath tumors were identified by the surgeon, which include pain, neurological deficits, schwannomatosis, and entrapment. Further, imaging characteristics, such as contrast enhancement, tissue edema, and border irregularity, were assessed. The degree of the presentation was assigned a score of -1 or +1, whether it suggested against or for malignancy, respectively. A 0 was assigned for an absence of the characteristic. This was retrospectively assigned to peripheral nerve sheath lesions diagnosed and treated from 2004 to 2015, and compared with pathology. Statistics were performed with 2-tailed t test and receiver operating characteristic (ROC) analysis.

RESULTS: A total of 50 tumors were identified, of which 8 were MPNST. The average score of benign lesions (0.1, range -1 to 1, SD = 0.58) was significantly different from the average score of MPNST (4.38, range 3-5, SD = 0.58) (P < .0001). A ROC analysis reveals an area under the curve (AUC) of 1.0, with a score of >2 having 100% sensitivity and specificity for malignancy.

CONCLUSION: Obtaining histopathological diagnosis is critical for the diagnosis of MPNST. The decision to operate may be difficult in cases with features suggestive of both benign and malignant pathologies. We demonstrate scoring criteria that can be quickly calculated on initial presentation, although validation with multiple surgeons and more patients would be required.