Conventional Direct Smear Yields Diagnostic Indicators of Gastric-Type Mucinous Carcinoma Compared with Cytomorphological Features Identified by Liquid-Based Cervical Cytology

Abstract

<b><i>Introduction:</i></b> Gastric-type mucinous carcinoma (GAS) of the uterine cervix is an adenocarcinoma subtype with a gastric phenotype that poses diagnostic pitfalls in cervical screening cytology because of its blunt morphologic atypia and the limited utility of human papillomavirus testing and ancillary immunochemical staining. Despite the recent widespread uptake of liquid-based cytology (LBC) systems, the cytomorphological features of GAS in LBC samples and the differential features between GAS and usual-type endocervical adenocarcinoma (UEA) remain unclear. <b><i>Methods:</i></b> Eight GAS cases, all of which were surgically treated following histological confirmation, were examined. Direct Papanicolaou-stained smears and LBC samples were reviewed and compared with 10 UEA cases as controls. Featured cytomorphological findings were as follows: background (mucinous, inflammatory, or necrotic), cell crowding (size of neoplastic cell clusters), cytoplasm (golden mucin and cell border), and nuclei (nuclear chromatin and nucleoli). <b><i>Results:</i></b> Of 18 adenocarcinomas, 16 were detected against a non-mucinous background in LBC samples, most of which were accompanied by mild to moderate inflammation. Clusters comprising &#x3e;300 neoplastic cells were identified in both GAS and UEA in conventional smears (CSs), while no LBC samples harboured clusters as large as these. Cell borders of GAS were more distinct than those of UEA in CSs (<i>p</i> &#x3c; 0.001), although fewer populations of neoplastic clusters revealed distinct cell borders in both GAS and UEA in LBC samples. Three of 8 and 2 of 8 GAS cases had golden mucin in CSs and in LBC samples, respectively, which was not detected in UEA at all. Nucleoli against fine nuclear chromatin were more pronounced in GAS than in UEA on CS (<i>p</i> = 0.03), although the difference between GAS and UEA was not apparent in LBC samples. <b><i>Discussion/Conclusion:</i></b> This study demonstrated that the diagnostic clues to detect GAS using the conventional approach, namely distinct cell borders and prominent nucleoli, are not useful for excluding UEA in LBC samples. Conventional cervical smears may indicate a diagnosis of GAS; however, specific high-risk HPV detection approaches, such as HPV test or immunocytochemical p16/Ki-67 dual staining, are desirable to differentiate GAS from UEA in the setting of LBC with ambiguous cytomorphological features.