p16INK4a and HPV E4 immunohistochemistry for the prediction of regression of cervical intraepithelial neoplasia grade 2-A historical cohort study.

Abstract

Cervical intraepithelial neoplasia grade 2 (CIN2) is a heterogeneous diagnosis with a high likelihood of spontaneous regression. Therefore, active surveillance for CIN2 has been implemented as an option in younger women in many countries. Yet, little is known about markers that may accurately predict the likelihood of regression to support active surveillance. Here, we aimed to assess whether p16^INK4a and HPV E4 status are associated with the likelihood of CIN2 regression. We conducted a historical cohort study on women aged 23-40 diagnosed with untreated CIN2 following cytology-based screening. Women were diagnosed at Aarhus University Hospital, Denmark from January 2000 to December 2010. Archived tissue samples were sectioned for p16^INK4a and HPV E4 immunohistochemistry and HPV genotyping. We used a modified Poisson model to estimate the relative risk of regression, adjusting for age and cytology (aRR). A total of 443 women with CIN2 were included. Most women (73.8%) were aged ≤30, and half had a high-grade cytology (48.8%). Overall, 47.6% regressed, and regression was less likely in p16^INK4a-positive compared to p16^INK4a-negative women (aRR 0.77; 95% CI 0.64-0.94). Among p16^INK4a-positive women, the risk of regression varied by HPV type and HPVE4 status, with lower risk in HPV16-positive women compared to those without (aRR 0.54; 95% CI 0.40-0.75) and in HPVE4-negative compared to HPVE4 positive women (aRR 0.73; 95% CI 0.54-0.98). When we restricted to expert-confirmed CIN2, the risk of regression did not vary by p16^INK4a or HPVE4 status, while HPV16 positive remained at a lower risk of regression compared to women without HPV16.