p16 Improves interobserver agreement in diagnosis of anal intraepithelial neoplasia

Evaluation of anal intraepithelial neoplasia (AIN) is subjective. Previous studies have shown p16 and Ki-67 expressions to correlate with AIN grade. Biomarkers like p16 and Ki-67 may improve interobserver agreement. The objectives were (1) to determine the extent of interobserver agreement in evaluating AIN on routine hematoxylin and eosin (H&E) sections and (2) to test whether p16 and/or Ki-67 staining improve interobserver diagnostic agreement.

Seventy-seven anal specimens were retrieved. Sections were stained with monoclonal antibodies against p16 and Ki-67. Blind to the original diagnoses, 4 pathologists assessed H&E alone, p16 alone, Ki-67 alone, and all 3 simultaneously. Diagnoses were normal/reactive, AIN I/HPV, AIN II, and AIN III. Agreement was calculated using kappa and S statistics.

Pathologists were board certified and had 2 to 25 years (mean = 13.6 years) of experience. Fair agreement was observed using H&E diagnosis alone (kappa = 0.38, S = 0.56). The p16 diagnostic evaluation demonstrated the highest agreement (kappa = 0.57, S = 0.73). Interobserver agreement for Ki-67 alone and for H&E/p16/Ki-67 combined were comparable to that of H&E alone (kappa = 0.4, S = 0.54 and kappa = 0.44, S = 0.62, respectively). When the pathologists' diagnoses for all diagnostic evaluations were compared with consensus diagnoses, the lowest average magnitude of disagreement was seen with Ki-67 alone, followed by p16 alone, H&E/p16/Ki-67 combined, and H&E alone.

Interobserver agreement for diagnosis of AIN was fair when based solely on H&E preparation. p16 alone improved interobserver agreement and demonstrated superior agreement when compared with H&E, Ki-67, and H&E/p16/Ki-67 combined.