Data sources: Medline, EMbase and Cochrane databases
Studies were included if they reported data on patients with a histologically-confirmed diagnosis of oral dysplasia. They also had to study at least one outcome measure and one intervention method or clinical risk factor. Outcome measures included malignant transformation rate (MTR) and time to malignant transformation (TMT). In studies on oral lesions, which contained a defined subset of patients with oral dysplasia, this subset but not all patients with oral lesions were included in the meta-analysis. Observational studies were included in the review due to the scarcity of randomised controlled trials with adequate follow-up period.
Data extraction and synthesis
Quality assessment was undertaken independently by two reviewers. When there was disagreement, a third reviewer was consulted. Quality criteria were agreed upon a priori, and the authors used a combination of quality assessment methods. Eligibility of studies was determined independently from the abstracts by two reviewers blinded to each other’s selections. Outcome data were abstracted independently by a researcher and a statistician and checked by a third reviewer. Heterogeneity was assessed graphically in a forest plot and a meta-analysis was conducted. Subgroup analysis was performed by histologic grade, clinical risk factors and treatment modality.
Fourteen non-randomised studies, reporting on 992 patients, were included. There was considerable heterogeneity between studies. The mean overall MTR was 12.1% (95% CI: 8.1%, 17.9%) and the mean TMT was 4.3 years. Histologic grade significantly affected mean MTR (p < 0.008). Furthermore, lesions that were not excised demonstrated considerably higher MTR than those that were excised (p=0.003)
Oral dysplasia showed a significant rate of transformation to cancer, which was related to grade, and was decreased significantly but not eliminated by excision. Findings suggest the need for surgical excision and continued surveillance, particularly in high-grade lesions.
Evidence-Based Dentistry 11, 91-92 (2010)