Risk-based screening turns a blind eye to many breast cancers in 40-something women
Screening mammography of women in their 40s based solely on risk as flagged by extremely dense breasts, very strong family history or both would lead to much higher mortality on the heels of more than 75 percent fewer screen-detected cancers in this group.
That’s from a study running in the December edition of the American Journal of Roentgenology.
Elissa Price, MD, of UC-San Francisco and colleagues retrospectively identified all cancers detected by screening mammography in women 40 to 49 years old at UCSF Medical Center between January 1997 and November 2012.
After excluding women with a personal history of breast cancer, they noted family history, breast density, type of malignancy, hormone receptor status and lymph node status.
From a field of 39,715 screening exams, the researchers recorded 136 cancers (cancer detection rate, 3.4 per 1,000).
Most tellingly, some 104 of the 136 patients—76 percent—had neither very strong family history nor extremely dense breasts.
Meanwhile, very strong family history was absent in the vast majority (119 of 136, or 87.5 percent), and extremely dense breast tissue was absent in like proportion (117 of 136, or 86 percent).
Among all three groups of low-risk patients, 79 percent had invasive cancers, of which 25 percent had axillary nodal involvement and 89 percent were estrogen receptor-positive.
Even looking at a risk-based strategy with an expanded definition of high risk—any first-degree family history or extremely dense tissue—the team found that 66 percent of malignancies would still be missed.
In their study discussion, Price et al. state that using very strong family history and/or extremely dense tissue as a foundation for risk-based screening strategies “would have failed to capture the great majority of cancers in our young population. These are women who develop consequential cancer that is often invasive and node-positive, but that is also likely to be treated successfully."
The study was first presented at RSNA 2014.