Breast cancer risk-stratification models can help radiologists identify women who may benefit from additional imaging exams as well as screening mammography and implementing other preventive strategies. An international, multi-institutional team of breast imaging specialists hypothesized that adding breast density to an established risk-stratification model would increase its accuracy. Their research using the Tyrer-Cuzick tool, published online May 11, 2019 in the Journal of Breast Imaging, confirmed their hypothesis.
The Tyrer-Cuzick model is used to calculate the likelihood of carrying the BRCA1 or BRCA2 mutations, and estimates the likelihood of a woman developing invasive breast cancer within 10 years and over the course of her lifetime. It estimates risk based on age, body mass index, age at menarche, obstetric history, age at menopause, family history, use of hormone replacement therapy, and history of ovarian cancer and/or benign breast conditions that increase breast cancer risk, such as hyperplasia or atypical hyperplasia. A recent British study from the Centre for Cancer Prevention of Queen Mary University in London determined that this long-term risk assessment tool has been accurate for at least 19 years.1
Led by Jennifer A. Harvey, MD, co-director of the breast care program at the University of Virginia School of Medicine in Charlottesville, the authors conducted a case-control study of 474 patients with breast cancer between 2003 and 2013 and an additional 2243 health control participants.2 All cancer patients had a full-field digital mammogram at the time of breast cancer diagnosis, and were between 40 and 79 years of age. The healthy controls were age-matched, and had at least one digital mammogram between 2003 and 2008.
The researchers reviewed mammograms from the contralateral breast taken before and closest to their cancer diagnosis. They measured breast density using an automated volumetric software program (Volpara®️Density™, Volpara Solutions, Wellington, NZ). The estimated percentage of the breast by volume occupied by fibroglandular tissue was defined as the primary breast density value, and the absolute fibroglandular and fat volume were secondary predictors. For the control group, the researchers reviewed mammograms performed closest to the time Tyrer-Cuzick questionnaires were completed. They also obtained B-RDS density categories from the clinical records of all participants.
They then developed a measure of breast density, defined as the difference between observed and expected density, independent from age at mammogram and BMI at the time the questionnaire was completed for the participants, and incorporated this measure into the Tyrer-Cuzick risk model.
The researchers reported that when the conventional Tyrer-Cuzick model was used, 4.8% of the study participants were identified as high risk. But when the volumetric breast density measure was added to the model, the risk increased to 6.8% of the participants, and up to 7.1% with the addition of BI-RADS density.
“Volumetric breast density and BI-RADS density add significant contributions to the predictive power of the Tyrer-Cuzick model,” the authors wrote, adding that “volumetric density has some practical advantages because it is fully automated and with excellent agreement with three dimensional (3D) magnetic resonance imaging (MRI).”
The authors also determined that both age and BMI were significantly negatively correlated with volumetric percent breast density, and that volumetric percent breast density decreased both with age and BMI. This was also true for BI-RADS density, but less so. BMI was significantly positively correlated with both fibroglandular and fat volume. Absolute fibroglandular volume was a weaker predictor compared with BI-RADS density or volumetric percentage density. The study also verified that women in the highest BI-RADS density category were three times more likely to develop breast cancer than women with predominantly fatty breasts.
They also verified that women in the highest Bi-RADS density category were three times more likely to develop breast cancer than were women with predominantly fatty breasts.
Breast density measures improve Tyrer-Cuzick breast cancer risk model. Appl Radiol.