Biomarkers in dysplasia of the oral cavity: A systematic review

Oral dysplasia is a potentially precancerous lesion diagnosed histologically. While the risk of progression is associated with histological grade, it is currently impossible to predict accurately which lesions will progress. More accurate markers predicting progression to cancer would enable the targeting of these lesions for more aggressive treatment and closer follow-up. We have performed a systematic review with pooling of data to assess the evidence for the use of biomarkers in predicting transformation of oral dysplasia into cancer. We systematically searched the Cochrane library, MEDLINE, EMBASE, AMED, Cinahl and the Kings Fund electronic databases using the terms: oral dysplasia, leukoplakia, erythroplakia, biomarkers and genetic markers. The following a priori selection criteria were used: longitudinal cohort or case-controlled studies of oral dysplasia that progressed to cancer. Cross-sectional studies and studies reporting only on leukoplakia were excluded. Data were extracted by two reviewers. Quality assessment was carried out using validated tools. We assessed the relative risk of progression form oral dysplasia to cancer and pooled data where possible. 2550 studies were identified, from which 288 were scrutinised in greater detail. Of these, 247 were excluded, mainly due to cross-sectional design. Of the 41 studies containing follow-up data, 28 were excluded, most commonly due to data only being available for lesions once they had progressed to cancer. A lack of clear histological definition of oral lesions was also a common finding. Data were extracted from 13 longitudinal studies. The evidence consists mainly of small, single centre, retrospective studies. In oral dysplasia, loss of heterozygosity (LOH), particularly at the 3p +/- 9p loci, increases the risk of progression to cancer (RR 17.60 (2.77, 108.37) p < 0.001), as does survivin (RR 30 (4.25, 197.73), p <= 0.001), matrix metalloproteinase (MMP 9), (RR 19.00 (1.56, 209.38) p = 0.02) and DNA content (RR 12.00 (1.17, 82.10) p = 0.03). Other markers identified by this review including p53, p73, MMP 1 and 2 and cathepsin L mRNA, did not predict progression. LOH, survivin, MMP 9 and DNA content are potential markers for increased risk of progression from oral dysplasia to cancer. Many methodological limitations have been identified by this review, however, and we recommend these results are interpreted with caution. Research into this field should concentrate on longitudinal design, with pooling of data from multiple centres to achieve larger cohorts. We recommend standardisation of definitions to allow appropriate comparisons to be made. (C) 2009 Published by Elsevier Ltd.