A recent study from UCLA has revealed that the risk evaluations for breast cancer in women vary significantly based on the risk assessment model utilized. As a result, recommendations provided to women are inconsistent, depending on the chosen model and the threshold used to categorize as “high-risk.”
The risk of developing cancer increases with age
Based on current data, it is estimated that approximately 12% of women born in the United States will experience breast cancer at some point in their lifetime. The probability of developing breast cancer tends to rise as one gets older.
With the advancement of precision medicine in healthcare, risk models for breast cancer have gained popularity in identifying women who could potentially benefit from medication to decrease breast cancer risk and complementary MRI screening. User-friendly risk models are conveniently accessible on the internet, and women usually receive a risk evaluation on their screening mammogram records. A crucial concern, though, is the precision level of these models.
Previously, a 5-year threshold of 1.67% was established. However, the Task Force advised a new, elevated 5-year threshold of 3%. Although current tools for evaluating breast cancer risk function effectively on a group level, they have received minimal attention in terms of how they perform for individuals, and there has been little consideration of the variance in risk evaluations for the 5-year threshold of ≥ 3.0% on an individual level.
Different models predict varying cancer risk
In the current study that evaluated over 31,115 women, researchers found that using the ≥ 1.67% threshold resulted in 21% of women being considered at high risk of breast cancer in the next five years but one model but were at average risk by another model. With the ≥3 thresholds, over 5% of women differed in risk severity among models. When all three models are used, around 46.6% of women are considered at high risk of breast cancer by a single model.
The study’s senior author Dr. Joann Elmore said that the study shows the risk of using a blanket approach to utilizing a risk prediction model in informing treatment decisions and medical screening.
