Diagnostic Accuracy of Oral Brush Biopsy Versus Scalpel Biopsy for Detecting Epithelial Dysplasia in Oral Potentially Malignant Disorders

Abstract

Introduction: Oral squamous cell carcinoma is often preceded by potentially malignant disorders, yet clinically distinguishing dysplastic or malignant lesions from benign mimics remains challenging, and routine biopsy of all suspicious lesions is impractical. The OralCDx® Brush Biopsy ('O' Test) has been proposed as a non-invasive adjunct for triaging such lesions, though its diagnostic reliability remains uncertain. This study evaluated the 'O' Test in predicting dysplasia or malignancy in clinically innocuous-looking oral potentially malignant disorders.

Aim and objective: To evaluate the diagnostic accuracy of the OralCDx® Brush Biopsy ('O' Test) in predicting epithelial dysplasia or malignancy in clinically innocuous-looking oral potentially malignant disorders, using scalpel histopathology as the reference standard.

Materials and Methods : In this prospective diagnostic study, 40 patients with clinically confirmed  oral potentially malignant lesions (>6 mm, unresponsive to empirical antimicrobial therapy) underwent both brush biopsy and concurrent scalpel/punch biopsy at the same lesion site. Diagnostic accuracy of the 'O' Test was assessed against histopathology using sensitivity, specificity, predictive values, McNemar's test, and Cohen's kappa; habit and demographic associations were assessed using the Chi-square test

Results: Of 40 patients (mean age 38.15 ± 9.45 years; 92.5% male), scalpel histopathology confirmed dysplasia in 20 (50%), whereas the 'O' Test identified dysplasia in only 5 (12.5%). The brush biopsy showed low sensitivity (25.0%; 95% CI 8.66–49.10%) but high specificity (100%; 95% CI 83.16–100%),with diagnostic accuracy of 62.5% and a false-negative rate of 75%.  McNemar's test showed significant discordance between the two methods (p < 0.001), with only fair agreement (Cohen's κ = 0.25).

Conclusion: The 'O' Test demonstrated poor sensitivity and an unacceptably high false-negative rate for detecting dysplasia in clinically innocuous oral lesions. Despite its non-invasive appeal, it cannot reliably substitute scalpel biopsy, which remains essential for definitive diagnosis.