Interobserver agreement in the reporting of cervical biopsy specimens obtained from women screened by visual inspection with acetic acid and hybrid capture 2.

Visual inspection with acetic acid (VIA) and human papillomavirus detection have sensitivity higher than cytology but lower specificity. The high false-positive rate of either test poses a challenge to the colposcopists who obtain biopsies from the innocuous changes and to the pathologists who have to interpret large numbers of specimens that are either normal or have low-grade abnormalities. Interobserver variation in histopathologic interpretations of cervical punch biopsy specimens is high, specially for the lower-grade abnormalities. Use of the modified Bethesda system to report histology in place of the cervical intraepithelial neoplasia (CIN) system has the potential to reduce such variability as there are fewer categories. The present study aimed to assess the interobserver agreement to interpret cervical punch biopsies when both pathologists followed the modified Bethesda classification system and also when one pathologist followed the modified Bethesda classification system and the other followed the CIN classification system. Colposcopy-directed punch biopsies were obtained from VIA and/or Hybrid Capture 2-positive women. The Institute pathologist interpreted the slides using the CIN system. Blinded review was done by 2 external pathologists who independently interpreted cervical punch biopsies using the Bethesda system. The Institute pathologist's diagnoses based on CIN system were converted post hoc into categories belonging to the Bethesda system for comparison. The overall agreement was poor (κ=0.36). The lowest agreement was observed in the low-grade squamous intraepithelial lesion category (κ=0.23) and the highest in the squamous cell carcinoma category (κ=0.76). The agreement between the reviewers, both of whom used the Bethesda system, was substantial.