Skip to content

Centers recently announced it will adopt ProFound AI® for Digital
Breast Tomosynthesis (DBT) throughout all of ImageCare’s mammography centers this
October as part of its “PINK Better Mammo” service. ProFound AI® was
the first artificial intelligence (AI) software for DBT, or 3D mammography, to
be FDA cleared; the software is clinically proven to improve accuracy and
efficiency for radiologists reading mammography.[i]

Breast Center, the women’s health services arm of ImageCare Centers, has been
using ProFound AI since September 2019. The technology has improved workflow
and reading accuracy at the facility since adoption, offering benefits to both
clinicians and patients. ProFound AI not only reduced the rate of false
positives and unnecessary recalls for women, it decreased the amount of
biopsies and increased the chance the biopsies performed were needed.

are always striving to adopt the latest in cutting-edge technology – we were
among the first in our communities to adopt DBT and ProFound AI,” according to
Lisa Sheppard, MD, founder of PINK Breast Center, located in Flemington and
Paterson, NJ. “As soon as we implemented ProFound AI, we started using it on
all of our DBT breast cancer screenings. It greatly improved our workflow and
enabled us to get back to rapid reads and offered the opportunity to provide
results for patients in real-time.”

or 3D mammography, offers many advantages over 2D mammography, including
increased cancer detection rates with fewer false positives that lead to
unnecessary and costly recalls, but it also increases reading time almost
two-fold, compared to 2D mammography alone. While 2D mammography typically
yields four images for each screening patient, DBT produces hundreds of images
for each patient, thus substantially increasing the daily workload for
clinicians. DBT is especially useful for women with dense breasts because
breast cancer and breast tissue both appear white on a mammogram, making it
difficult for the radiologist to read. DBT improves the radiologist’s ability
to find cancers and can reduce the need for biopsies. 

Built with the latest in deep-learning
technology, ProFound AI rapidly analyzes each DBT image, or slice, detecting
malignant soft tissue densities and calcifications with unrivaled accuracy.
Certainty of Finding and Case Scores are assigned to each detection and each
case respectively. These scores represent the algorithm’s confidence that a
detection or case is malignant, providing crucial information for radiologists
that may assist them in clinical decision making.

“ProFound AI is revolutionizing the way mammography
is read. With superior performance and sensitivity, the software offers
unrivaled accuracy and time-savings benefits,” according to Stacey Stevens,
President of iCAD, the manufacturer of ProFound AI. “As one of the latest tools
in the fight against breast cancer, we are pleased that it will soon be
available to more women in New Jersey.”

AI for DBT was clinically proven in a large reader study to increase radiologists’
sensitivity by 8 percent, minimize the rate of false positives and unnecessary
recalls by 7 percent and reduce reading time by 52.7 percent.i ProFound AI was also clinically proven
to slash reading time by up to 57.4 percent for radiologists reading cases of
women with dense breasts.[ii]

ProFound AI highlights an area, we know it’s something to investigate. It’s
much more selective than other CAD technologies and offers a remarkable improvement
in terms of the focus”, says Dr. Sheppard. She adds, “This technology has made
a tremendous impact on patient care at PINK Breast Center. It helps to ensure
that the biopsies we perform are more likely to be cancer, and we also now have
fewer false positives and callbacks.”


[i] Conant,
E et al. (2019) Improving Accuracy and Efficiency with Concurrent Use of
Artificial Intelligence for Digital Breast Tomosynthesis. Radiology: Artificial
Intelligence. 1(4). Accessed via

[ii] Hoffmeister,
J. (2018). Artificial Intelligence for Digital Breast Tomosynthesis – Reader
Study Results. [White paper]. Accessed via