Accuracy of preoperative clinical examination and imaging for mandibular resection in oral cancer: a systematic review of diagnostic test accuracy studies.

The aim of this systematic review is not only to analyse the accuracy of clinical examination and radiological preoperative assessment of mandibular invasion reported in isolation, but to highlight those reports that have combined them. A total of 1636 titles and abstracts published between 1995 - 2000 were screened following a literature search in PubMed. Keywords were "mandible" and "squamous cell carcinoma". A total of 90 full manuscripts were reviewed with 24 meeting defined inclusion/exclusion criteria and yielding the data reported. The most sensitive test was single photon emission tomography with eight out of the 10 studies reporting sensitivity higher than 95%. Magnetic resonance imaging (MRI) demonstrated superior sensitivity but was less specific than computed tomography (CT). A single report attempted to report the combined CT and MRI scans with a separate expert reporting but did not result in more reliable detection. Periosteal stripping was not reported, and there was insufficient data to establish the value of new technologies. This review confirms that, to our knowledge, there are no reliable data on the results of combining imaging techniques with or without clinical examination. It emphasises the lack of data for the combination of preoperative techniques to enhance safe oncological resection of the mandible. Based on the evidence gathered in this review an algorithm of assessment of possible mandibular invasion is proposed. With new technologies available and 3-dimensional models to help plan the mandibular resection and reconstruction, the potential of combining preoperative investigations should be fully realised through prospective research.