A team of researchers from the University of Chicago Medicine has developed innovative polygenic risk score (PRS) models that significantly enhance the prediction of breast cancer risk among women of African ancestry. This advancement aims to address the higher breast cancer mortality rates in this demographic, which have been attributed to several factors, including existing genetic models that inadequately assess risk for these populations.
Addressing Disparities in Breast Cancer Risk Prediction
Despite progress in genetic testing, women of African ancestry continue to experience disproportionately high death rates from breast cancer. Traditional risk prediction models have primarily been based on data from white women of European ancestry, resulting in less accurate risk assessments for African American women, especially concerning aggressive tumor types like triple-negative breast cancer (TNBC).
The new PRS models were developed using genetic data from over 36,000 women, making them the most comprehensive tools for predicting breast cancer risk in this historically underserved population. The findings are detailed in a study published in Nature Genetics.
“Polygenic risk scores worked well for European-Americans but weren’t accurate for African American women due to smaller sample sizes and greater genetic diversity,” said Dezheng Huo, Ph.D., Professor of Public Health Sciences and senior author of the study. By collaborating with researchers from 20 institutions, the team aimed to improve prediction accuracy.
Developing Tailored Models for Improved Accuracy
The research utilized data from the African Ancestry Breast Cancer Genetics Consortium, which included women from the U.S., the Caribbean, and Sub-Saharan Africa. The team created specific PRS models for various breast cancer types, including overall breast cancer, estrogen receptor positive (ER+), estrogen receptor negative (ER-), and TNBC.
The effectiveness of each model was evaluated using its area under the curve (AUC) score, which indicates how well a model can differentiate between individuals who will develop breast cancer and those who will not. The new models achieved AUC scores ranging from 0.61 to 0.64, a notable improvement over previous models that scored between 0.56 and 0.58.
To make the models more accessible and cost-effective, the team also developed simplified versions. For instance, a TNBC model utilizing only 162 genetic markers demonstrated comparable performance with an AUC of 0.626, facilitating clinical use.
“With improved risk prediction, doctors can start screening earlier for women at higher risk and tailor care based on a woman’s specific risk profile,” Huo noted. The study revealed that women in the highest 1% of risk scores faced a lifetime risk of 25.7% for developing breast cancer, while those with the top risk score for TNBC had a lifetime risk of 7.4%.
These findings suggest that high-risk women could benefit from early screening starting as young as age 32, compared to the current guidelines recommending screening at ages 40 or 45.
Enhancing Predictions with Family History
Integrating family history into the new PRS models further strengthens risk predictions. Women in the top 1% of PRS scores with a first-degree relative diagnosed with breast cancer had a lifetime risk exceeding 50%. This level of risk warrants earlier and more frequent screening, along with potential preventive measures such as medications or genetic counseling.
The accuracy of the new models was validated across several independent datasets, including the All of Us study and additional studies involving women of African ancestry. In one instance, the TNBC model achieved an AUC of 0.652, confirming the reliability of these new tools across diverse populations.
While the study primarily focused on African American women and those of West African ancestry, the researchers emphasized the necessity for further exploration. Genetic distinctions among West, East, North, and South African populations, as well as the broader global African population, warrant continued research.
“These advanced testing models bring us closer to a future where everyone, regardless of their ancestry, has an equal opportunity for early detection, effective treatment, and improved survival rates,” Huo concluded.
The full study, titled “Improved polygenic risk prediction models for breast cancer subtypes in women of African ancestry,” will be published in March 2026.
