Accuracy of MRI in prediction of tumour thickness and nodal stage in oral squamous cell carcinoma

C T Lwin, R Hanlon, D Lowe, J S Brown, J A Woolgar, A Triantafyllou, Simon N Rogers, F Bekiroglu, H Lewis-Jones, H Wieshmann, R J Shaw

    Research output: Contribution to journalArticle (journal)peer-review

    66 Citations (Scopus)

    Abstract

    We aim to compare radiological with histological tumour thickness (RTT with HTT) for oral squamous cell carcinoma (OSCC), and the ability of both to predict cervical metastasis. The MRI images and histopathology reports of 102 consecutive OSCC cases were compared and the relationship between RTT and HTT, calculated as a ‘‘shrinkage factor’’ by the gradient of the best fitting regression line. Most (69%) tumours appeared thicker on MRI than was revealed by histopathology. Shrinkage factor was 0.70 (interquartile range 0.63–0.77, correlation co-efficient 0.63) for all cases, 0.87 (IQR 0.80–0.95, CC 0.88) for tongue and 0.65 (IQR 0.49–0.82, CC 0.45) for floor of mouth sub-sites. RTT did not correlate well with the presence of nodal metastases in any sub-site, i.e. there was no clinically applicable cut-off value of RTT to determine the prescription of elective neck dissection. Although RTT has some predictable relationship with HTT, this varies between sub-sites with tongue the most accurately predicted shrinkage using axial MRI. It is not possible from either the MRI staging of neck or tumour thickness to safely determine the need for neck dissection in OSCC. It is necessary to re-evaluate the benefit of MRI as a staging investigation (particularly for early stage OSCC) and further explore the contribution of molecular biomarkers and ultrasound.
    Original languageEnglish
    Pages (from-to)149-154
    JournalOral Oncology
    Volume48
    Issue number2
    DOIs
    Publication statusPublished - 2012

    Fingerprint

    Dive into the research topics of 'Accuracy of MRI in prediction of tumour thickness and nodal stage in oral squamous cell carcinoma'. Together they form a unique fingerprint.

    Cite this