Interobserver reproducibility in determining p16 overexpression in cervical lesions: Use of a combined scoring method

Abstract

Objectives: To evaluate interobserver reproducibility of a combined scoring method for immunohistochemical interpretation of p16 overexpression in cervical lesions. Materials and methods: p16 immunostaining was performed in cervical samples from 183 patients, including 69 normal, 42 low grade squamous intraepithelial lesions(LSIL), 36 high grade SIL (HSIL), and 36 squamous cell carcinomas(SCCAs). Each case was evaluated by a combined scoring method based on the percentage of positive cells (score 0-3), the intensitiy of staining (score 0-3), and the distribution pattern (score 0-2). Immunoexpression for p16 was considered as positive when the combined score was 4-8 and negative with a score of 0-3. Ten pathologists with varied experience in interpretating p16 immunostains evaluated each slide independently. Results: All normal cervical squamous epithelia (69/69) were uniformly negative for p16. All HSILs (36/36), all SCCAs (100/100), and all but one of the LSILs (40/41, 97.62%) showed positive expression. In 172 of 183 cases (93.99%), p16 interpretation was concordant with all pathologists. Eleven cases with discordant results included 10 LSILs and 1 normal mucosa sample. Percentage of agreement of each pathologist pair ranged from 96.7-100% (mean 98.1 ± 0.96%) with mean kappa value of 0.96 ± 0.0201 (range 0.93-1.000). Conclusion: The proposed combined scoring method shows good reproducibility among the participating pathologists and good correlation with the histologic diagnosis. This method may be a useful guide in the interpretation of p16 expression in cervical epithelial lesions.