Anal intraepithelial neoplasia: correlation of grade with p16INK4a immunohistochemistry and HPV in situ hybridization.

Accurate diagnosis and grading of anal intraepithelial neoplasia (AIN) can be problematic, especially in separating AIN from anal transitional-zone epithelium. Immunohistochemistry (IHC) for p16 correlates with human papillomavirus (HPV) integration into the host genome, and HPV subtyping by in situ hybridization (ISH) is now readily available. To investigate if p16 would help in more accurately diagnosing and grading AIN, particularly when attempting to distinguish benign transitional-zone epithelium from high-grade AIN, we separately assessed these stains in a blinded manner on a large number of consecutive anal biopsies and anal tissues and correlated the findings with the diagnosis and grade of AIN. One hundred thirty-three consecutive anal tissue specimens, from 128 patients were studied. One hundred and eight were anal biopsies and 25 were hemorrhoidectomy specimens. All specimens were stained with hematoxylin and eosin, p16 (Lab Vision), and HPV-ISH (Ventana HPV-III Inform). No gold standard was chosen, rather, comparisons between the 3 tests were made and agreement between tests was tested for statistical significance using the kappa value (kappa). The comparisons included AIN grade (negative, 1, 2, 3) with nuclear intensity of p16 IHC (0 to 3+), AIN grade with IHC nuclear staining patterns (contiguous, patchy/rare), AIN grade with HPV-ISH [negative, low-risk (LR), high-risk (HR)] and, nuclear intensity of p16 IHC with HPV-ISH. One hundred percent of AIN2 and 3 cases were strongly positive for nuclear p16, whereas 80% of AIN1 were positive. Yet, 33% of AIN negative cases were positive for nuclear p16, although with less nuclear intensity than for AIN2 or 3. The kappa value for AIN/nuclear p16 intensity agreement was 0.61 (ie, substantial agreement). Seventy-nine percent of AIN3 cases were strongly positive for HR, and 9% for LR HPV; whereas 75% of AIN2 were positive for HR and 5% for LR HPV. Forty percent of AIN1 cases were positive for HR and 24% for LR HPV. Yet, 12.5% of AIN negative cases were positive for HPV for both LR and HR types. The kappa value for AIN/HPV agreement was 0.62 (ie, substantial agreement). One hundred percent of HR and 85% of LR HPV were positive for p16. Yet, 6% of negative p16 were positive for HPV (all were LR), whereas 43% of negative HPV showed some p16. Of the latter, 40% were 3+ by p16. The kappa value for HPV/nuclear p16 intensity agreement was 0.56 (ie, moderate agreement). Of interest, 30 cases were negative for AIN and p16 staining and of these, 2 (7%) were positive for HPV (both LR subtype). Three cases positive for HR HPV were negative for AIN with only patchy nuclear p16 positivity. We conclude that the correlation between AIN and p16 and HPV is strong enough to be quite useful in distinguishing true AIN from benign mimics, such as benign transitional-zone epithelium. In selected cases, p16 and HPV may be valuable in grading as well.