RESUMEN
INTRODUCTION: histologic interpretation of hematoxylin and eosin-stained cervical biopsies is subject to substantial discordance among pathologists. Immunohistochemical staining for p16INK4a can reduce inter-observer disagreement. We did a cross-sectional study to evaluate the utility of p16INK4a staining in the assessment of cervical biopsies in Nairobi, Kenya. METHODS: hematoxylin and eosin-stained sections from 91 colposcopic biopsies diagnosed as negative for dysplasia or as cervical intraepithelial neoplasia (CIN) grade 1-3 from 2011-2013 in Nairobi, Kenya, were reviewed and immunostained for p16INK4a. Agreement in interpretation of cervical biopsies was compared between primary and consensus review results. RESULTS: on primary evaluation, 16 cases were negative for squamous dysplasia; 23 were CIN 1; 37 CIN 2; and 15 CIN 3. On consensus review, 32 cases were negative for dysplasia; 19 were CIN 1; 16 CIN 2 and 24 CIN 3. Agreement was moderate between primary and consensus histology review results for the diagnosis of low-grade versus high-grade squamous intraepithelial lesions (Kappa = 0.568). None of the cases negative for dysplasia were positive for p16INK4a expression, but in primary and consensus review results, 17% and 5% cases of CIN 1; 49% and 69% of CIN 2, and 80% and 96% of CIN 3 were p16INK4a positive, respectively. CONCLUSION: there was significant variability in the interpretation of cervical biopsies on hematoxylin and eosin between primary and consensus review assessments. 75% of CIN 1 cases that were upgraded to CIN 2 during consensus review expressed p16INK4a. These findings demonstrate the role of p16INK4a in increasing diagnostic accuracy and as a marker of high-grade CIN 2/3.