Clinicopathological and prognostic significance of immunophenotypic characterization of endocervical adenocarcinoma using CLDN18, CDH17, and PAX8 in association with HPV status.

ABSTRACT

In 2020, the WHO published a new system for classifying invasive endocervical adenocarcinoma based on histological features and high-risk human papillomavirus (HPV) infection. However, immunophenotypes of each histological subtype require further investigation. We immunohistochemically analyzed 66 invasive endocervical adenocarcinomas using three cell-lineage-specific markers claudin 18 (CLDN18) for gastric, cadherin 17 (CDH17) for intestinal, and PAX8 for Müllerian epithelial cells. We identified five immunophenotypes of endocervical adenocarcinoma gastric (21%); intestinal (14%); gastrointestinal (11%); Müllerian (35%); and not otherwise specified (NOS) (20%). Adenocarcinomas with gastric immunophenotype, characterized by aging (p = 0.0050), infrequent HPV infection (p < 0.0001), concurrent lobular endocervical glandular hyperplasia (p = 0.0060), lymphovascular invasion (p = 0.0073), advanced clinical stage (p = 0.0001), and the poorest progression-free (p < 0.0001) and overall (p = 0.0023) survivals, were morphologically compatible with gastric-type adenocarcinoma of the WHO 2020 classification. Conversely, most adenocarcinomas with Müllerian (91%) and intestinal (89%) immunophenotypes were HPV associated and morphologically compatible with usual- or intestinal-type adenocarcinomas of the WHO 2020 classification. The morphology of adenocarcinomas with gastrointestinal immunophenotype was intermediate or mixed between those of gastric and intestinal immunophenotypes; 57% were HPV associated. Adenocarcinomas with NOS immunophenotype were mainly HPV associated (85%) and histologically poorly differentiated. Multivariate analysis revealed that gastric (p = 0.008), intestinal + gastrointestinal (p = 0.0103), and NOS (p = 0.009) immunophenotypes were independent predictors of progression-free survival. Immunophenotypes characterized by CLDN18, CDH17, and PAX8 exhibited clinicopathological relevance and may improve the diagnostic accuracy and prognostic value of conventional histological classification.