Stathmin-1 expression as a complement to p16 helps identify high-grade cervical intraepithelial neoplasia with...

A fundamental controversy in using biomarkers to diagnose cervical cancer precursors [ie, squamous intraepithelial lesions (SIL)] is their lack of specificity for high-grade SIL [HSIL; cervical intraepithelial neoplasia grade 2/3 (CIN2/3)]. Stathmin-1 (STMN), a microtubule-destabilizing protein important in mitosis, is overexpressed in a variety of malignancies including those from the Müllerian tract and may be associated with poor outcome. However, the use of STMN as a diagnostic marker in cervical SILs has not been explored. A total of 193 cervical samples, including benign cervix and squamous and glandular lesions, were evaluated for STMN, p27 (a known cell cycle regulator inversely expressed with STMN in normal and many neoplastic tissues), p16, and Ki67 expression by immunohistochemical analysis. SILs were independently scored and classified as benign, low-grade SIL (LSIL/CIN1), and HSIL (separated into CIN2 or CIN3) on the basis of a majority diagnosis. Each diagnosis was correlated with biomarker expression independent of hematoxylin and eosin review. STMN was normally expressed in basal cells of the ectocervix and was absent in benign endocervix;"positive" STMN staining was defined as immunoreactivity in at least two thirds of the epithelial thickness. A majority hematoxylin and eosin diagnosis was obtained in 189/193 cases on initial independent review, with squamous epithelia ultimately classified as benign, CIN1, CIN2, and CIN3 in 25, 56, 11, and 16 cases, respectively. STMN was positive in 5/56 (9%) CIN1, 5/11 (45%) CIN2, 14/15 (93%) CIN3, all adenocarcinoma in situ (19/19), and all invasive squamous cell carcinoma (32/32) and adenocarcinoma (34/34) cases. In contrast, 40/56 (71%) CIN1, 11/11 (100%) CIN2, and 15/16 (94%) CIN3 lesions were p16 positive (defined as diffuse block staining in at least one third of the epithelial thickness). One CIN3 was negative for both p16 and STMN. Ki67 and p27 staining was variable in squamous lesions. These results demonstrate that STMN is overexpressed in virtually all cervical carcinomas and CIN3 lesions. In contrast to p16, which often stains LSIL and a small subset of reactive biopsies, STMN has greater specificity for CIN3 and has the potential to distinguish these latter lesions from the majority of low-grade precursors and negative/reactive cervical biopsies. STMN overexpression appears independent of proliferation, maturation, and human papilloma virus cytopathic effect and may be useful diagnostically in identifying HSIL. Further studies correlating STMN staining with lesion persistence or morphologic progression, in the context of CIN grade, are warranted.