P16 and Ki67 immunostaining is a useful adjunct in the assessment of biopsies for HPV-associated anal intraepithelial neoplasia.
P16 is a tumor suppressor gene product, shown to be overexpressed in most cervical carcinomas and dysplasias associated with high-risk human papilloma virus (HPV) infection. HPV is also associated with anal squamous dysplasias and carcinomas. Significant interobserver and intraobserver variation exists in the interpretation of biopsies for anal intraepithelial neoplasia (AIN). This study was undertaken to assess the potential role of p16 and Ki67 immunohistochemical expression in refining the diagnosis and grading of AIN.One-hundred and four anal biopsies from 74 patients were retrieved from the surgical pathology files of the department. After discrepancies were resolved and concurrence was achieved by at least 2 of 3 reviewing pathologists, the diagnoses were as follows: 37 negative, 12 condylomas without overt dysplasia, 14 AIN I, 25 AIN II, and 16 AIN III. p16 and Ki67 expression was evaluated by ABC immunoperoxidase staining whereas the presence of the high-risk subtypes of HPV virus was determined by in situ hybridization on a subset of the biopsies. Results were reviewed by 2 pathologists and positive and negative staining was correlated with H&E diagnoses. Nuclear and/or nuclear and cytoplasmic staining was considered as positive for p16 when present in >10% of squamous cells. Two patterns of positive p16 staining were observed: (1) "spotty" in which positive cells were scattered throughout the lesion and (2) "band" in which >90% of contiguous cells in the lesion stained positive. A band-like pattern of p16 immunoreactivity was seen in 21.4% AIN I, 80% AIN II, and 87.5% AIN III cases. None of the condylomas and only 1 of the negative cases showed a band of p16 positive staining. Spotty p16 immunoreactivity was observed in 8.1% negative, 8.3% condyloma, 14.3% AIN I, 12.0% AIN II, and 12.5% AIN III cases. More than 50% of nuclei stained positive for Ki67 in 28.6% AIN I, 48.0% AIN II, and 75.0% AIN III cases but in none of the negative or condyloma cases. On the basis of these results, a band-like pattern of p16 staining and Ki67 positivity in >50% of the squamous cell nuclei were strongly associated with high-grade AIN. Conversely, absence of a p16 band of positivity coupled with Ki67 positivity in <50% of nuclei was frequently associated with benign lesions. Band like p16 staining also correlated strongly with the presence of high-risk HPV. Most AIN I lesions stained similar to the nondysplastic cases. A small subset of biopsies studied did not conform to the pattern described above: 4 of 14 (28.6%) AIN I lesions showed a band-like pattern of p16 staining and/or >50% Ki67 positive nuclei. Two of these cases were positive for high-risk HPV DNA. 4 of 25 (16.0%) AIN II lesions comprising 9.8% of the 41 high-grade AINs (AIN II and III) showed spotty p16 positivity and <50% Ki67 positive nuclei. One was positive for high-risk HPV DNA. We conclude that when used together and evaluated in conjunction with H&E stained sections, p16 and Ki67 immunoexpression is a useful adjunct in the diagnosis and grading of AIN.