The immunohistochemical expression of p16 in formalin-fixed and paraffin-embedded histological sections was evaluated in a retrospective study comprising a low-grade group of 100 cases of cervical intraepithelial neoplasia (CIN) 1, a high-grade group of 50 cases of CIN 2 to 3, and a benign group of 50 cases of normal tissue or benign lesions in the uterine cervix. The cases were consecutive within each group and had a minimum follow-up period of 5 years. Positive reaction for p16 was detected in all cases in the high-grade group and in only 3 cases in the benign group. In the low-grade group, a total of 9 cases had to be excluded. The remaining 91 cases in the low-grade group showed positive reaction for p16 in 65 cases (71%), including 23 cases that progressed to a high-grade lesion, 36 cases that revealed normal cytological and/or histological picture during the follow-up period, and 6 cases that persisted as CIN 1. A total of 26 cases (29%) in the low-grade group showed negative reaction for p16. All but one of these p16 negative cases in the low-grade group had a benign or normal outcome. This case showed a high-grade lesion in the follow-up period and was probably a high-grade lesion from the beginning and so underestimated as CIN 1. These results reveal that the negative predictive value of p16 to predict the outcome of the cases of CIN 1 is as high as 96%, which strongly suggest an important role of p16 in the assessment of this type of lesion.

The quality of cervical histopathology is critical to cervical cancer prevention, cancer treatment, and research programs. On the basis of the histology results further patient management is determined. However, the diagnostic interpretation of histologic hematoxylin-eosin (H&E)-stained slides is affected by substantial rates of discordance among pathologists. Overexpression of the cyclin-dependent kinase inhibitor p16INK4a, a cell cycle regulating protein, has been shown to be strongly correlated with dysplastic lesions of the cervix uteri. In this study, we assessed whether p16INK4a immunohistochemistry may increase the performance of pathologists in diagnosing squamous lesions in cervical punch and cone biopsies. When using a consecutive p16INK4a-stained slide in conjunction to the H&E-stained slide, interobserver agreement between 6 pathologists improved significantly for both cervical punch and cone biopsies (P < 0.001). For punch biopsies (n = 247), kappa value increased from 0.49 (moderate agreement) to 0.64 indicating substantial agreement, and interobserver agreement for cone biopsies (n = 249) improved from 0.63 (conventional H&E slide reading) to 0.70 when H&E-stained slides were read conjunctively with p16INK4a-stained slides. In comparison to a common consensus diagnosis established by 3 independent experts, 4 pathologists reached an improvement with the conjunctive p16INK4a test, 2 of them showing significantly better agreement (P < 0.001 and P = 0.002, respectively), p16INK4a immunohistochemistry as an adjunct to conventional H&E-stained specimens thus contributes to a more reproducible diagnosis of cervical intraepithelial neoplasia and may be a valuable aid for the interpretation of cervical histology.