In honor of Breast Cancer Awareness Month, Stacey Stevens, president of iCAD, sits down with AXIS Imaging News to discuss the importance of knowing one’s risk of breast cancer and how modern technologies can lead to better outcomes for women.

AXIS Imaging News: What new technology has iCAD recently introduced?

Stacey Stevens: We’ve recently launched the latest generation of ProFound AI® Risk, which is now available for digital breast tomosynthesis (DBT), or 3D mammography. ProFound AI Risk is the world’s first commercially available clinical decision support tool that provides an accurate short-term breast cancer risk estimation that is truly personalized for each woman, based only on information garnered from a 2D or 3D mammogram, along with additional factors including age and breast density.

This technology is designed to aid physicians in optimizing individualized screening to determine the appropriate utilization of clinical resources and improve outcomes for patients. It uniquely combines a range of risk factors to offer superior performance and accuracy in assessing short-term risk compared to traditional, commonly used breast cancer risk models. It is an easy-to-integrate model that empowers physicians to truly personalize each patient’s screening and to find breast cancer early, when it may be more easily treated.

AXIS: How does this technology work?

Stevens: The ProFound AI Risk score considers breast complexity, including textures, patterns, and shapes of breast tissue; suspicious masses and calcifications and their distribution in the left and right breast; breast density and the asymmetry of density; age; and ethnicity, if available, to calculate the woman’s short-term, absolute risk of breast cancer. All this information is within a woman’s screening mammogram, making risk assessment simple. The score is automatically derived from the mammography exam, making implementation easy, with no questionnaires, portals, or staff required to implement it.

Results include the short-term (one, two, or three-year, aligned with that country’s screening program) breast cancer risk category (low, general, moderate, and high). Specifically designed to factor in racial and ethnic backgrounds, ProFound AI Risk offers an equitable and inclusive approach to precision screening. The algorithm also factors in clinically relevant global screening guidelines and more than 15 country incidence and mortality reference tables, for alignment with that country’s general population.

AXIS: What clinical evidence supports this technology?

Stevens: The latest version of ProFound AI Risk is supported by an internal validation study led by researchers at the Karolinska Institutet in Stockholm, Sweden, as well as iCAD scientists. The DBT image-based model reached an area under the curve (AUC) of 0.83 (95% CI 0.80,0.85).

Moreover, the previous version of ProFound AI Risk is supported by a recent study led by Karolinska Institutet researchers based on the prospective screening cohort Karolinska Mammography Project for Risk Prediction of Breast Cancer (KARMA), which took place from 2011-2017. The initial version of ProFound AI Risk, the FFDM image-based model reached an AUC of 0.73 (95% CI 0.71,0.74). Researchers found a statistically significant superior AUC performance when comparing the ProFound AI Risk model to the Tyrer-Cusick v8 model (+11 points) or GAIL model (+12 points) at two years, all including density within their calculations.

High-risk women, as determined by ProFound AI Risk, were more likely diagnosed with stage II and >= 20 mm tumors and less likely with stage I and estrogen receptor-positive tumors. Researchers concluded ProFound AI Risk identified women at high likelihood of being diagnosed with breast cancer within two years of a negative screen and in possible need of supplemental screening.

AXIS: What’s the benefit of a short-term risk estimation?

Stevens: Although traditional long-term risk models can be valuable for prevention, and in some markets are required for MRI coverage, short-term risk models based on mammograms can identify women in immediate need of supplemental screening. By identifying a woman’s short-term risk of developing cancer, we believe ProFound AI Risk can revolutionize personalized screening, leading to improved efficiencies and outcomes as well as reduced harms and costs.

AXIS: How will this technology benefit patients?

Stevens: Breast cancer recently surpassed lung cancer as the No.1 diagnosed cancer in the U.S., excluding non-melanoma skin cancers. Additionally, the National Cancer Institute recently predicted that there could be almost 10,000 excess deaths from breast and colorectal cancers over the next decade as a direct result of delayed screening due to the pandemic. It’s now more important than ever for clinicians to personalize screening regimens for women and to detect cancers earlier, when they may be more treatable.

Most mammography screening programs are not individualized, so in order to efficiently screen for breast cancer in patients, the individual risk of the disease must be determined. Physicians have traditionally estimated risk by examining known risk factors, such as family history, but about 85% of breast cancers occur in women who have no family history of breast cancer.

So, unless a woman has a known family history or a personal history of intervention, most women simply do not know their risk. With the help of ProFound AI Risk, clinicians will be able to be more accurate in early cancer detection, which we know can have a tremendous impact on women—from treatment to outcomes.


  • Eriksson M, Czene K, Strand F, Zackrisson S, Lindholm P, Lång K, Förnvik D, Sartor H, Mavaddat N, Easton D, Hall P. Identification of Women at High Risk of Breast Cancer Who Need Supplemental Screening. Radiology. 2020
  • Breast feature complexity findings from ProFound AI includes subtle radiomic breast tissue features such as texture, pattern and shapes.
  • Age risk factor utilizes country-specific Incidence and Mortality tables (I&M) within the iCARE package as a part of the ProFound AI Risk absolute score calculation. Over 15 country I&M tables are available for wide ancestry/country applicability. [1] U.S. Breast Cancer Statistics. Accessed via