CT-scan is a valuable tool to detect mandibular involvement in oral cancer patients.

In patients with oral squamous cell carcinomas (OSSC) it is desirable to avoid unnecessary bone resection without neglecting the overall surgical treatment goal of tumor-free margins. Whereas computed tomography (CT) is most commonly used to detect mandibular invasion, there are conflicting reports regarding the accuracy of CT. Therefore, the aim of this study was to reinvestigate the accuracy of CT in predicting mandibular involvement by OSSC. One hundred and seven patients with OSSC who received a mandibulectomy were included. Before treatment all patients underwent a contrast-enhanced multi-detector CT. Axial 3 or 1.25 mm thick images were reconstructed for evaluation in overlapping technique and displayed in a bone (1400/400 HU) and a soft tissue window (350/50 HU). CT scans were examined by three investigators and compared with the histological findings. The radiological examination showed a high interrater reliability (Cronbachs alpha 0.982). Comparing the radiological findings with the histological results the CT showed 8 false-positive results and 8 false-negative patients. The quality criteria for detecting bone involvement of OSSC by CT were calculated as follows: sensitivity 82.6%; specificity 86.9%; positive predictive value 82.6%; negative predictive value 86.9%. However, in all false-positive patients a sagittal bone defect of 15.1mm could be found presumably caused by pressure of the tumor, but no histologically detectable bone infiltration. Modern CT (1-2 mm sections) is a valuable tool for surgical treatment planning. If bone invasion is detected, a mandibulectomy seems always reasonable. In radiologically negative cases histological assessment is necessary to detect mandibular involvement.