Background: Grading of oral epithelial dysplasia (OED) by a pathologist is currently the key guide used for treatment planning of oral potentially malignant disorders (OPMDs). Conventional oral examination (COE) clinically detects OPMDs but may not predict their risk status to transform to cancer. Therefore, there is a need for a reliable test to predict OED in OPMDs. Aim: This study was conducted to evaluate COE, liquid based cytology (Cytopath) and DNA image cytometry (Ploidy) in predicting OED in OPMDs. Methods: A total of 179 patients from Malaysia, India and Indonesia underwent COE followed by brush biopsies and scalpel biopsies. Brush-biopsy samples were analyzed for cytopathology and DNA ploidy at Dental Faculty, University of Malaya. Histopathological findings of presence/absence of OED were used as the reference standard. Calculations for sensitivity (Sn), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV) and accuracy (A) were done for individual tools and in combinations. The Youden index (Sn+Sp-1) was used as a measure of overall performance. The relevant medical ethics committees of the different research locations approved the study. Results: For COE, the sensitivity (Sn) was high (100%) and the specificity (Sp) was low (5.9%), while both Cytopath and Ploidy showed a low sensitivity (Sn) (28.6% and 22.2%) and high specificity (Sp) (94.3% and 82.3%). All 3 tools individually have high negative predictive value (NPV) for predicting presence of OED (COE-100%, Cytopath-66.7%, Ploidy-78.5%). When combining outcomes from all 3 tools, the best performance indicated by Youden index (42.1) is which defines a positive case when both COE and Cytopath show abnormal. In general, using results from at least 2 tools had better Youden indices than using these tools individually. Conclusion: COE as a screening tool by virtue of its high Sn would be a suitable first level diagnostic test, while the Cytopath and the Ploidy individually with high Sp may be used as a second level test to predict presence of OED. Combining the COE with cytopathology would be the best combination for a high performance of the tools. Cytopathology (when performed by a trained cytologist) would allow for most of the false positives from the first level test to be correctly identified as true negative at the second level. Longitudinal data are needed to assess which of these may correctly identify the malignant potential of OPMDs. Acknowledgment: Grant: High Impact Research - Ministry of Higher Education (HIR-MOHE UM000025/C3)
Predicting the clinical outcome of oral potentially malignant disorders using transcriptomic-based molecular pathology