Investigating the clinical significance of EGFR expression using machine learning in a series of Iraqi patients with triple-negative breast cancer.

Abstract

Breast cancer is a heterogeneous disease with a distinct profile of the expression of each tumor. Triple-negative breast cancer (TNBC) is a molecular subtype of breast cancer characterized by an aggressive clinical behavior linked to loss or reduced expression of estrogen, progesterone, and Her2/neu receptors. The study's main objective was to investigate the clinical significance of epidermal growth factor receptor (EGFR) overexpression in a series of Iraqi patients with TNBC. The sectional analytic study involved immunohistochemical analysis of EGFR expression in randomly selected 53 formalin fixed paraffin embedded tissue blocks of TNBC cases out of 127 Iraqi patients with TNBC and correlated expression data with clinicopathological parameters including survival time. Machine learning (statistical tests and principal component analysis (PCA)) was used to predict the outcome of the patients using EGFR expression data together with clinicopathological parameters. EGFR was expressed in approximately 28% of TNBC cases. We estimated the risk of mortality and distant metastasis based on EGFR expression and clinicopathologic factors using the principal component analysis (PCA) model. We found a substantial positive correlation between clinical stage and distant metastasis, clinical stage and death, death and distant metastasis, and death and positive EGFR expression. Overall, EGFR expression was linked to a poor prognosis and increased mortality. A higher risk of distant metastasis and death was associated with an advanced clinical stage of the tumor. Furthermore, the existence of distant metastases increased the risk of death. These findings raise the possibility of using EGFR expression data with other clinicopathological parameters to predict the outcome of patients with TNBC.