ABSTRACT
Cytokeratin 15 (CK15) has been described as a stem cell marker in human organs and its expression is seen in breast tissue. CK15 expression is associated with aggressive features in endometrial and oesophageal cancers, but data on the breast are lacking. This study aims to investigate the clinicopathological associations and prognostic significance of CK15 in breast carcinomas. A multi-institute cohort of breast carcinomas were retrieved. Clinicopathological and outcome data were obtained and compared with immunohistochemical expression CK15 and a panel of biomarkers. In total, 1,476 cases were included, with an expression rate of 3.5%, preferentially expressed in luminal subtypes (p=0.024), with luminal B carcinomas being the highest (4.7%), as opposed to basal-like (1%) and HER2-overexpressed carcinomas (0%). Except for nodal stage (p=0.013) and nodal metastasis (p=0.048), oestrogen (p=0.035) and progesterone receptor (p=0.001) positivity, there were no associations with other clinicopathological parameters. A trend was observed with shorter breast cancer specific survival (BCSS) in CK15-positive luminal B carcinomas (p=0.062). On further subgroup multivariate analysis of luminal B HER2-negative carcinomas, CK15 expression exhibited robust correlation with shorter BCSS (HR=9.004, p=0.001) and disease-free survival (HR=7.085, p<0.001). Restricted to luminal breast carcinomas, specifically luminal B HER2-negative, CK15 is demonstrated to be a robust independent predictor of higher risk of recurrence and shorter survival, with potential as a clinical prognostic marker and an exclusive stem cell marker for this subgroup of carcinomas.
ABSTRACT
AIMS: Trichorhinophalangeal syndrome-1 (TRPS1) has been proposed as a novel breast marker with equally high expression in breast cancer (BC) subtypes, making it a useful diagnostic tool. Here, its expression was evaluated alongside other commonly used markers [GATA3, GCDFP15, mammaglobin (MGB) and SOX10] in a large cohort of BCs (n = 1852) and their corresponding nodal metastases. Its usefulness as a diagnostic tool and its correlation with clinicopathological features were assessed. METHODS AND RESULTS: TRPS1 was expressed at 75.8% overall in the BC cohort, with at least 58% expression among BC subtypes. It was less sensitive than GATA3 for luminal and HER2-overexpressing (HER2-OE) cancers (luminal A: 82 versus 97%; luminal B: 80 versus 95%; HER2-OE: 62 versus 76%), but it was the most sensitive for TNBC (60 versus ≤ 41%). It showed a stable expression in nodal metastases (primary tumour 76 versus nodal metastasis 78%), unlike a reduced nodal expression for GATA3 (86 versus 77%). TRPS1 outperformed GATA3 in detecting non-luminal cancers when paired with other breast markers. TRPS1 and GCDFP15 was the most sensitive combination in TNBC detection, with a 76% detection rate. For TRPS1-negative and GCDFP15-negative TNBCs, SOX10 was more sensitive than GATA3 (29 versus 24%). CONCLUSIONS: TRPS1 is a highly sensitive marker for all breast cancer subtypes, outperforming GATA3 in non-luminal cancers and displaying the highest sensitivity for TNBC detection when combined with GCDFP15. It is a valuable addition to the breast marker panel for accurate identification of BC.