Interobserver and intraobserver bias exists in the interpretation of anal dysplasia

INTRODUCTION: Natural history of progression from anal intraepithelial neoplasia to invasive carcinoma remains unproven. The risk of progression may be linked to the severity of dysplasia. Important therapeutic decisions are thus based on the severity of anal intraepithelial neoplasia. Consistency and reliability in the interpretation of anal intraepithelial neoplasia are unproven. METHODS: One hundred ninety anal biopsy specimens were identified for review of dysplasia with a six-point grade system from normal to invasive cancer, evidence of human papillomavirus infection, and quality of histology by three pathologists. RESULTS: Results revealed poor to moderate agreement on grading of quality of histology (weighted kappa score, 0.07-0.22), human papillomavirus status (weighted kappa score, 0.24-0.53), and dysplasia (weighted kappa score, 0.38-0.7). Complete agreement between the original pathology and the three pathologists was observed in only 32 percent of cases. Analysis of 86 slides previously read by one of the pathologists revealed only moderate agreement, with a weighted kappa score of 0.64. CONCLUSION: Significant interobserver and intraobserver bias exists in the interpretation of anal intraepithelial neoplasia. These inconsistencies may explain the uncertainty about the natural progression of anal intraepithelial neoplasia and the varied results of surgery reported for anal intraepithelial neoplasia in the literature.