Sanmitra Iwanski, ScreenPoint Medical

How Digital Breast Tomosynthesis Revolutionizes Mammography

Do you take care of yourself? Sure, you eat right and you exercise, but do you get your screenings? 

Screening mammography is an evidence-based, proven tool to improve outcomes for women through early breast cancer detection and treatment. In fact, breast cancer detected in the earliest stages is generally survivable. Furthermore, despite some confusing and contradictory recommendations over the last 10 years, compliance with screening mammography is quite high in the US due to advocacy efforts, and continues to gain traction in other regions of the world. 

Despite all its value, screening mammography has a few key limitations. 

  1. Information overload that impedes workflow. 
  2. Decreased confidence in the densest breast tissue. 


Before the advent of digital breast tomosynthesis (DBT or 3D mammography), mammograms were already some of the largest datasets processed in the radiology department. The advent of increased information through DBT came with a tsunami of data that the radiologist has to read… Whereas a 2D mammography study may have consisted of as few as 20 million pixels1 of data, a 3D mammography study can contain 4.5 BILLION pixels of data2. The smallest cluster of microcalcifications associated with cancer can be about 9 pixels of data in these studies.

Radiologists are looking for a needle in a haystack… 

When Transpara’s DBT algorithm was introduced, the reader study submitted to the FDA showed that radiologists could save a substantial amount of time per DBT study when using Transpara to read3. Furthermore, as it is labeled for concurrent reading, that means that radiologists can use the Transpara report as a roadmap for areas of concern, helping find the most concerning regions quickly and effectively. The greatest value of Transpara comes in UNMARKED studies (Low risk studies). In the aforementioned reader study, about 7 in 10 studies were classified as low risk (no marks) and saved readers 26% of their time per study.


Too often, workflow and confidence are perceived as a tradeoff. If a radiologist spends infinite time on a mammogram, can every lesion be found? Those results don’t exist. But what does exist is a backlog of images, all of which need to be read, and half of which are from women with dense breast tissue. Density has been a topic of conversation in the mammography world for decades, but only recently has there been a surge to action around density. Fibroglandular tissue, which causes the appearance of density on mammographic images, is also the type of tissue in which cancer is likely to develop. And cancer and fibroglandular tissue are both radio-opaque, appearing white to the human eye on a mammogram. As a result, radiologist sensitivity drops from over 80% in the fattiest/least dense breasts to around 50-60% in the most dense breasts. A drop in sensitivity can cause cancers obscured during screening that become symptomatic before they are found. In a field focused on early detection, this combination is deadly for some women, and stressful for the radiologists who care for them. 

Despite overwhelming demands, radiologists do an amazing job of finding breast cancer. And they do even better with Transpara support! In a 2023 paper from Koch, et al, radiologists reading with Transpara experienced no decline in sensitivity regardless of breast density.

Researchers at UCLA have conducted extensive research to better understand how Transpara can help with a diverse, complex patient population. In a study presented at RSNA 2023 by Dr. Tiffany Yu, 150 women who presented with interval cancers had their prior screening mammogram analyzed using Transpara. Of these 150 screening mammograms, 35 were classified as having minimal but actionable signs of cancer. Transpara detected 89% of those lesions. This study demonstrates that readers WITH the support of Transpara can confidently detect subtle areas of concern. 

Ultimately, when choosing software, radiologists need to trust that the software will make their lives better. By delivering both confidence AND workflow enhancements, Transpara supports reading effectiveness and enables radiologists to spend more time caring for patients.

  1. GE Senographe Pristina, in 2D acquisition mode (ScreenPoint internal data) ↩︎
  2. Hologic Selenia Dimensions Clarity, in 3D acquisition mode with synthetic 2D (ScreenPoint internal data) ↩︎
  3. Van Winkel, et al ↩︎

Explore the Latest AI Insights, Trends and Research

Sanmitra Iwanski, ScreenPoint Medical