Clinical and morphological factors associated with two groups of concordant breast cancer immunophenotypes

Abstract

This study aimed to describe the prevalence of invasive breast cancer (IBC) in women assisted in a public hospital in Brazil and to establish a correlation between two models of classification by immunophenotypes, one of them based on the 13th St. Gallen Conference classification and the other on biomarker-defined subtypes based on HER2 and oestrogen receptor status, as described in World Health Organization (WHO). We selected IBC of 1335 cases between 1994 and 2018. Univariate frequencies and associations were estimated using chi-square tests. The concordance between the two immunohistochemical analysis models above mentioned using Cohen's kappa coefficients. The most prevalent subtype was luminal B/HER2, and the frequency of tumours with a worse prognosis was 62.7%. Has been identified an association between histological grade 3 (G3) and the worst prognostic subtypes: non-luminal A, Triple Negative Breast Cancer (TNBC), non-ER+/HER2- and ER-/HER2-. A similar association was found in nuclear G3 tumours. The results showed agreement between 99.48% and 100% when we compared the two immunohistochemical analysis models. Furthermore, there was an absolute agreement between the two models of immunohistochemical analysis. These results can contribute to institutions that do not have a molecular investigation, enabling accessible routine practice tools. Among the most important questions addressed in this work is the association between histological and nuclear G3 with the worst prognostic subtypes. Using the St. Gallen Conference classification and HER2 and ER status based on subtypes verified the feasibility of selecting IBC with different prognoses and correlated them with recognized predictive and prognostic factors.