Doctor examines mammogram of breast on monitors

Image source: Adobe Stock/okrasiuk 

News • Mammogram analysis

AI accurately classifies breast density

An artificial intelligence (AI) tool can accurately and consistently classify breast density on mammograms, according to a new study.

Breast density reflects the amount of fibroglandular tissue in the breast commonly seen on mammograms. High breast density is an independent breast cancer risk factor, and its masking effect of underlying lesions reduces the sensitivity of mammography. Consequently, many U.S. states have laws requiring that women with dense breasts be notified after a mammogram, so that they can choose to undergo supplementary tests to improve cancer detection. 

In clinical practice, breast density is visually assessed on two-view mammograms, most commonly with the American College of Radiology Breast Imaging-Reporting and Data System (BI-RADS) four-category scale, ranging from Category A for almost entirely fatty breasts to Category D for extremely dense. The system has limitations, as visual classification is prone to inter- and intra-observer variability. To overcome this variability, researchers in Italy developed software for breast density classification based on deep learning with convolutional neural networks, a sophisticated type of AI able to discern subtle patterns in images beyond the capabilities of the human eye. The researchers trained the software, known as TRACE4BDensity, under the supervision of seven experienced radiologists who independently visually assessed 760 mammographic images. 

The results were published in Radiology: Artificial Intelligence.

Selection of mammographic mediolateral oblique views of breasts with different breast density from women between 51 and 68 years of age. (A–D) Examples of human readers (HR)-(AI) agreement for category a (68 years), b (66 years), c (51 years), and d (54 years); B shows an example of a breast with a benign mass. (E–H) Examples of HR-AI disagreement; E was classified as a by HR, and as b by AI (67 years); F was classified as b by HR, and as a by AI (68 years); G was classified as c by HR, and as d by AI (55 years); H was classified as d by HR, and as c by AI (52 years). Note - Breast Imaging Reporting and Data System (BI-RADS): category a (almost entirely fatty), category b (scattered fibroglandular), category c (heterogeneously dense), category d (extremely dense).

Image source: Magni et al, Radiology: Artificial Intelligence; @RSNA 2022

External validation of the tool was performed by the three radiologists closest to the consensus on a dataset of 384 mammographic images obtained from a different center. 

TRACE4BDensity showed 89% accuracy in distinguishing between low density (BI-RADS categories A and B) and high density (BI-RADS categories C and D) breast tissue, with an agreement of 90% between the tool and the three readers. All disagreements were in adjacent BI-RADS categories. “The particular value of this tool is the possibility to overcome the suboptimal reproducibility of visual human density classification that limits its practical usability,” said study co-author Sergio Papa, MD, from the Centro Diagnostico Italiano in Milan. “To have a robust tool that proposes the density assignment in a standardized fashion may help a lot in decision making. Such a tool would be particularly valuable, the researchers said, as breast cancer screening becomes more personalized, with density assessment accounting for one important factor in risk stratification.” 

“A tool such as TRACE4BDensity can help us advise women with dense breasts to have, after a negative mammogram, supplemental screening with ultrasound, MRI or contrast-enhanced mammography,” said study co-author Francesco Sardanelli, MD, from the Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Donato in San Donato, Italy. 

The researchers plan additional studies to better understand the full capabilities of the software. “We would like to further assess the AI tool TRACE4BDensity, particularly in countries where regulations on women density is not active, by evaluating the usefulness of such tool for radiologists and patients,” said study co-author Christian Salvatore, PhD, senior researcher, Scuola Universitaria Superiore Pavia, and co-founder and chief executive officer of DeepTrace Technologies. 

Source: Radiological Society of North America


Read all latest stories

Related articles


News • Breast cancer screening

Mammography: AI on par with human readers, study finds

Mammographic screening is used to detect breast cancer, but is prone to false findings. Could an AI reader help improve this? UK researchers compared the performance of man and machine.


Article • Supplemental imaging

The next breast screening advancement: Contrast-enhanced mammography

With the recent recommendation changes from the European Council in 2022, how radiologists screen for breast cancer is changing. Mammography has long been an essential technology in screening for…


News • Deep learning CAD in sonography

Computer-aided diagnosis improves breast ultrasound expertise

A new study suggests that deep learning–based computer-aided diagnosis (CAD) for breast lesion classification on ultrasound can significantly improve radiologists' diagnostic performance.

Related products

teamplay Mammo Dashboard

Business Intelligence

Siemens Healthineers · teamplay Mammo Dashboard

Siemens Healthcare GmbH

Artificial Intelligence

ScreenPoint Medical · Transpara

ScreenPoint Medical
50° wide-angle Tomosynthesis


Siemens Healthineers · 50° wide-angle Tomosynthesis

Siemens Healthcare GmbH
Advanced Breast-CT - nu:view

Mammo CT

AB-CT · Advanced Breast-CT - nu:view

AB-CT – Advanced Breast-CT GmbH

Digital Mammography

Drtech Europe · Aidia

AI-Rad Companion

Artificial Intelligence

Siemens Healthineers · AI-Rad Companion

Siemens Healthcare GmbH
Subscribe to Newsletter