Risk model fails to accurately predict breast cancer in women with atypia

July 12, 2010

The Tyrer-Cuzick risk model was no better than flipping a coin in predicting which women with atypical hyperplasia would develop invasive breast cancer within ten years, according to a study of 331 women followed for an average of 14 years.  The study, published online by the Journal of Clinical Oncology, was conducted at the Mayo Clinic and involved a subset of women from the Mayo Benign Breast Disease cohort, a group of 9,376 women who had benign breast biopsies between 1967 and 1991.

Atypical hyperplasia is an accumulation of abnormal cells in a breast duct or lobule and is a risk factor for development of invasive breast cancer.  Investigators predicted that the Tyrer-Cuzick risk model, which estimates the risk of breast cancer using information about BRCA1/2 mutation carrier status and other risk factors including age at menarche, age at first childbirth, and body mass index, would perform better than the Gail model in predicting breast cancer risk in those with atypcial hyperplasia, since it includes more extensive information about family history and benign breast disease.

However, the Tyrer-Cuzick model over predicted the development of breast cancer by almost two to one, predicting that 59 of the 331 women with atypia would develop invasive breast cancer, while only 31 actually did within ten years. The Gail Model was no better at predicting an individual's risk of developing invasive breast cancer.  The investigators looked to see if the Tyrer-Cuzick model would fare any better in women with a strong family history of breast cancer, and found the model still fared poorly.

The investigators suggest it may be better for women to consider the absolute risk of developing breast cancer rather than the results from risk models, when making any decisions on preventive efforts.  In this study group of women with atypia, the 10-year absolute risk of developing invasive breast cancer was 9.2%.