Immunostaining for carcinoembryonic antigen does not discriminate for early recurrence in breast cancer. The ECOG experience

Carcinoembryonic antigen (CEA) immunohistochemistry was evaluated by 11 surgical pathologists with sections from 147 postmenopausal women with node‐positive breast cancer. Carcinoembryonic antigen staining in breast cancer tissues has been correlated with a worse prognosis. This association was studied with a clinically characterized population of Eastern Cooperative Oncology Group (ECOG) patients using precisely the peroxidase‐antiperoxidase methodology which had been employed in another published study. In 50% of the cases, the study pathologists were uncertain whether CEA was or was not present in the cancers. Various groupings of the pathologists' interpretation were compared with the observed disease‐free intervals in the patients. These analyses suggested no association of perceived CEA staining with the biological course of the cancers. Two reference pathologists who examined the sections in a similar way also gave nonprognostic interpretations. There is no convincing evidence that pathologists can reliably interpret the CEA content in the same breast cancer tissue sections. There is no observed correlation between immunohistochemical evidence of CEA in a breast cancer tissue section and the biological behavior of that cancer.