Distinction of endocervical and endometrial adenocarcinomas: immunohistochemical p16 expression correlated with human papillomavirus (HPV) DNA detection.

Determining the origin of uterine adenocarcinomas can be difficult in biopsy and curettage specimens because the morphologic spectrum of endocervical and endometrial adenocarcinomas overlaps. In hysterectomy specimens, the primary site is often equivocal for tumors that involve the lower uterine segment and endocervix and lack identifiable precursor lesions. Most endocervical adenocarcinomas (ECAs) contain high-risk human papillomavirus (HPV) DNA, whereas endometrial adenocarcinomas (EMAs) rarely do. p16 is an inhibitor ofcyclin-dependent kinases, and overexpression of p16 has been observed in cervical intraepithelial lesions and invasive carcinomas associated with high-risk HPV types. We evaluated the utility of immunohistochemistry for p16 in the distinction of ECAs and EMAs. p16 expression was assessed in 24 unequivocal EMAs and 19 unequivocal ECAs and correlated with HPV DNA detection by in situ hybridization and polymerase chain reaction. These assays were then used to assist in the classification of four lower uterine segment/upper endocervical adenocarcinomas (LUS/EC-A) of equivocal origin. p16 expression was moderate-strong and diffuse in 18 ECAs (median 90% of tumor cells positive, range 90%-100%), and weak and diffuse in one. Fourteen of these were positive for HPV DNA, whereas 5 lacked detectable HPV DNA by in situ hybridization; one of these 5 was positive by polymerase chain reaction. In contrast, EMAs displayed weaker staining with patchy distribution (median 30% of tumor cells positive, range 5%-70%) and none contained HPV DNA by in situ hybridization. Two LUS/EC-As, which were positive for HPV, exhibited strong, diffuse p16 expression, consistent with endocervical origin of the tumors. The remaining 2 LUS/EC-As showed patchy p16 staining and did not contain detectable HPV DNA, consistent with the endometrial origin of the tumors. The p16 expression pattern can distinguish ECAs from EMAs. Compared with HPV DNA detection by in situ hybridization, p16 immunohistochemistry appears to be a more sensitive and easier to perform method for distinguishing ECAs from EMAs, can be used to assist in the classification of LUS/EC-As of equivocal origin, and should be evaluated for its utility in the prospective classification of uterine adenocarcinomas in curettage specimens prior to hysterectomy.