Effectiveness of anatomic criteria for predicting parotid tumour location.

Abstract

We evaluated the accuracy, positive predictive value (PPV), negative predictive value (NPV), specificity and sensitivity of eight anatomic landmarks to differentiate parotid deep lobe tumours from superficial lobe tumours: the lateral margin of the retromandibular vein (RMV), a straight line from the facial nerve trunk (FN trunk) to the mandibular ramus (FN line), a straight line from the FN trunk to the RMV (tRMV), a straight line from the FN trunk to the lateral margin of the masseter (tMasseter), a straight line from the ipsilateral vertebral posterior end to the RMV (U-line), an arc with a radius of 8.5mm centred on the mandibular ramus (Conn's arc), a straight line from the lateral surface of the masseter muscle to the lateral margin of the RMV (rMasseter) and an angle from the FN line, tumour and the lateral margin of the masseter muscle (FTM angle). A total of 181 patients with a parotid gland tumour who underwent parotidectomy at a tertiary hospital were identified retrospectively from May 2005 to May 2013. Pre-operative computed tomography and intraoperative findings were compared to evaluate each landmark. rMasseter (accuracy: 85.5%, PPV: 90.0%, NPV: 85.12%, specificity: 98.1%, sensitivity: 22.2%) and tMasseter (accuracy: 86.3%, PPV: 80.0%, NPV: 87.1%, specificity: 97.1%, sensitivity: 44.4%) showed superior results as diagnostic criteria. rMasseter and tMasseter were useful as anatomic landmarks to differentiate a parotid deep lobe tumour from a superficial lobe tumour.