Inter- and intra-observer variability in three grading systems for oral epithelial dysplasia

Area of Science:

• Oral pathology
• Histopathology
• Cancer diagnostics

Background:

• Oral epithelial dysplasia (OED) grading is crucial for predicting malignant potential of oral premalignant lesions.
• Existing OED grading systems are subjective and exhibit variability.
• Assessing inter- and intra-observer variability is key to improving diagnostic reliability.

Purpose of the Study:

• To evaluate the inter- and intra-observer variability of three OED grading systems: Binary, WHO (2005), and Ljubljana.
• To compare the diagnostic consistency across different grading systems for oral leukoplakia.
• To identify potential improvements for reducing subjectivity in OED diagnosis.

Main Methods:

• Histopathological grading of 63 leukoplakia cases by three pathologists using Binary, WHO (2005), and Ljubljana systems.
• Pathologists were blinded to case identity and grading system used.
• Inter- and intra-observer variability assessed using multivariate kappa analysis over repeated observations.

Main Results:

• Inter-observer agreement was slight to poor across all three systems (Binary, WHO, Ljubljana).
• Intra-observer agreement varied from slight to moderate across the systems.
• The binary system demonstrated the best overall inter- and intra-observer agreement among the evaluated systems.

Conclusions:

• The binary classification system offers improved inter- and intra-observer agreement for OED grading.
• Enhanced intra-observer agreement was observed across all systems and individual parameters.
• Objective parameter definition, reproducible system structuring, and pathologist training are recommended to minimize diagnostic variability.