Background: Establishing the risk of malignant transformation (MT) in oral leukoplakia is usually based on grading oral epithelial dysplasia (OED) on biopsy tissue, for which two systems are proposed: a 3-tier and a binary system. Only very few actuarial studies have tested the accuracy of such methods in predicting MT, especially for the binary system. This study aimed to assess the accuracy of the two grading systems in predicting MT in a cohort of oral leukoplakia (OL) from Brazil, with follow-up data.
Methods: The sample comprised 878 individuals diagnosed with OL from 2005 to 2018. Follow-up data were obtained both locally and from the regional cancer registry. All lesions were graded using both the 3-tier and the binary systems. Kaplan-Meier curves (Log-rank Mantel-Cox) were used to assess risk and kappa to assess interobserver agreement.
Results: Thirty-five individuals underwent MT (4%). Both systems demonstrated prognostic value, though the 3-tier system proved superior, with OR 9.23 (3.42-23.69), PPV 0.152, NPV 0.98, compared to binary OR 3.49 (1.79-6.79), PPV 0.079, NPV 0.976. Interobserver agreement was also superior in the 3-tier system (0.47, p < 0.05) compared to the binary system (0.139, p = 0.39). Combining the two systems enhanced prognostic values (OR 14.28, PPV 0.217, NPV 0.981).
Conclusion: The 3-tier system presented superior prognostic value to the binary system. Combining both systems to double-grade intermediate lesions might enhance risk assessment.
Keywords: leukoplakia; oral cancer; risk analysis; risk factor; risk scores.
© 2022 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.