Statistical Profiling of Oral Cancer and the Prediction of Outcome

Statistical Profiling of Oral Cancer and the Prediction of Outcome:

Abstract

Background

The global burden of oral squamous cell carcinoma (OSCC) remains formidable. Identifying factors predictive of aggressive tumour behaviour, disease progression and reduced survival time may assist in early identification of ‘high‐risk’ patients and appropriately target combination cancer therapies.

Methods

A retrospective review of 467 OSCC patients treated over a 19‐year period facilitated detailed clinico‐pathological database analysis and determination of clinical outcome categories based upon time to progressive disease (loco‐regional tumour recurrence and/or distant metastasis), overall death and OSCC‐related death (death directly attributable to OSCC). Odds Ratio (OR) and Hazard Ratio (HR) statistical measures were used to investigate relationships between patient demographics and clinico‐pathological tumour features with clinical outcome.

Results

Older age at presentation (p=0.002) and a history of previous non‐head and neck cancer (p=0.010) increased the risk of overall death. OR for progressive disease development (p = 0.008) and OSCC‐related death (p = 0.019) were most significant for buccal tumours. HR confirmed advanced stage disease increased the risk of progressive disease (p<0.001), overall death (p<0.001) and OSCC‐related death (p<0.001). Positive resection margins were associated with a higher risk of OSCC‐related death (p=0.023). Significantly lower risks for progressive disease development (p=0.002) and OSCC‐related death (p=0.012) were seen in patients undergoing neck dissection, whilst combination chemo‐radiotherapy reduced HR for overall death (p<0.001) and OSCC‐related death (p=0.011).

Conclusion

Statistical profiling of OSCC clinico‐pathological data identifies significant influences on clinical outcome. This study adds evidence to the hypothesis that buccal SCC displays aggressive tumour behaviour and poor clinical outcome.