Background: This work evaluated algorithmic bias on biomarkers using electronic pathology reports of female breast cancer across five subgroups: National Cancer Institute Surveillance, Epidemiology, ...