Skip to content Provide accessibility feedback
For Providers
For Patients
Customer Stories
About Us
Contact Us
Magnifying glass Search

How Automated, Objective Density Measures Impact Risk Assessment: Research Highlights from RSNA 2021

Volpara Health – Published on December 1, 2021

Volpara Health AI-powered breast software for optimizing breast density, risk assessment and positioning highlighted at this year’s conference

CHICAGO, DECEMBER 1, 2021 — Automated, objective breast density measures in risk assessment and tumor characteristic prediction was an important topic for Volpara Health at the 2021 Annual Radiological Society of North America (RSNA) Meeting. A total of 11 abstracts highlighted the use of Volpara software in this regard, as well as in studies related to mammography positioning, deep learning models, and optimizing breast cancer screening.

Two studies by Stamatia Destounis, MD and colleagues investigated the association of automated mammographic density (MD) and breast tumor characteristics (TCs). In the paper “Breast Density Associations with Tumor Characteristics Among Screen-detected and Interval Breast Cancers Diagnosed in a US Screening Setting,” MD was assessed by both BI-RADS® and Volpara software for a population of 318 screen-detected breast cancers and 100 interval cancers. Volpara software measured volumetric breast density (VBD) as well as providing a Volpara Density Grade (VDG). The findings showed that both VBD and VDG had stronger associations with TCs than BI-RADS. This indicates that using volumetric MD measures could help triage women with dense breasts to supplemental screening, potentially reducing the impact of MD and TCs associated with poorer outcomes.

The second study from Destounis et al., “Volumetric Breast Density is a Strong Predictor of Breast Tumor Size in a US Screening Population,” found that women with VDG 4 breasts had 3.8 times the risk of being diagnosed with tumor sizes greater than 2 cm compared to women with VDG 1 and 2 combined. Compared to non-dense patients, BI-RADS 4 was not significantly associated with tumor size.

In the study, “Impact On Risk Categorization With Inclusion Of Mammographic Density In The Tyrer-Cuzick Model”, Dr Destounis and colleagues retrospectively re-calculated the Tyrer-Cuzick version 7 risk scores of 59,257 patients using version 8 of the Tyrer-Cuzick model. Using Volpara’s VBD as the density input resulted in 11.4% of patients being considered high risk, compared to only 6.9% of patients when using BI-RADS as the density input, showing the potential value of volumetric density in identifying high-risk women who may benefit from supplemental screening.

On the topic of mammography quality and positioning, the study “Association Of Automated Breast Positioning Assessment with Technical Repeats and Recalls in the Nationalized U.K. Screening Program,” presented the association between mammogram images repeated due to inadequate positioning and automated PGMI scoring for breast positioning. PGMI stands for Perfect, Good, Moderate, Inadequate—indicating the quality of the mammogram based on the presence or absence of certain key positioning metrics. The study found that automated breast positioning assessment showed significant differences in PGMI scores between the accepted exams, technical repeats, and technical recalls, where the set of images reviewed included 1,340 accepted exams, 2,134 technical repeats, and 144 technical recalls. With these results in mind, automated image quality assessment algorithms can provide objective and timely feedback, with the potential to reduce technical recalls and help facilities continually improve image quality to meet MQSA standards.

Volpara’s AI-based software was also used in the following studies presented at this year’s RSNA:

  • “Denser or Just Thinner? Mammographic Breast Density Characterization Around the World”
  • “Deep Learning Predicts Interval and Screen-detected Cancer from Negative Screening Mammograms: A Case-case-control Study In 6369 Women”
  • “Triaging Women From MRI To Mammography To Adapt Screening To Changes In Breast Density Using Artificial Intelligence”
  • “AI-Triaging Of Breast MRI For Radiological Review In The Screening Of Women With Extremely Dense Breasts”
  • “Deep-libra: An Open-Source Artificial-intelligence Method For Robust Quantification Of Breast Density With Independent Validation In Breast Cancer Risk Assessment”
  • “Replacing A Radiologist By AI In Dutch Population Based Breast Cancer Screening And The Impact Of Breast Density On Performance”
  • “Commercially Available AI System For Breast Cancer Detection Shows Promise For Risk Prediction, Including Among Women With Dense Breasts”

About Volpara Health

Volpara provides an advanced AI software platform that works with a healthcare provider’s expertise to enable a high-quality, optimized, and personalized cancer screening experience. From the time a patient enters a clinic to the moment they obtain key results, the Volpara Breast Health Platform collects and analyzes information to better understand a patient’s breast cancer risk, while also objectively evaluating image quality and workflow-improvement opportunities. These capabilities are being extended to lung cancer screening. The Volpara Breast Health Platform is supported by numerous patents, trademarks, and regulatory registrations including FDA clearance and CE marking, and is validated by a volume of peer-reviewed publications unrivaled in the breast screening industry.


Media Contact:
Chris K. Joseph
Volpara Health