New Research Finds iCAD’s ProFound AI Risk for Digital Breast Tomosynthesis is 2.4x More Accurate than Traditional Lifetime Risk Models
Using U.S. guidelines, ProFound AI Risk found
14% of women studied who had a negative screen had
almost 20 times higher risk of developing breast cancer in the next year than
the general risk population
iCAD recently announced that promising clinical
research supporting ProFound AI® Risk for Digital Breast Tomosynthesis (DBT) – the world’s first clinical
decision support tool that provides an accurate short-term breast cancer risk
estimation based on age, breast density and mammographic features – was recently
published in the peer-reviewed journal, Science TranslationalMedicine.[i]
In the study, which involved 154,200 women screened at four
participating U.S. screening sites[ii]
using DBT from 2014-2019, researchers at the Karolinska Institutet in
Stockholm, Sweden found ProFound AI Risk accurately determined women who were
at a higher risk of developing breast cancer, with an area under the receiver
operating characteristics curve (AUC) of 0.82i, (95% CI 0.79-0.85). AUC is a standard performance measurement for examination
procedures that incorporates sensitivity and specificity into a single metric
of overall performance. This data supports iCAD’s internal research, which
previously found ProFound AI Risk for DBT offers an AUC of 0.80 (95% CI 0.76,
0.83).[iii]
“iCAD’s Breast AI Suite offers a complete
portfolio of unrivaled breast cancer detection, density assessment, and
short-term risk evaluation AI solutions, and we are pleased to see this
compelling data further validate the clinical value of ProFound AI Risk, the
latest addition to our breast AI portfolio,” said Stacey Stevens, President and
CEO of iCAD, Inc. “Physicians have traditionally estimated breast cancer risk
by examining the patient’s known risk factors, such as family history, but
about 85% of breast cancers occur in women who have no family history of breast
cancer.[iv]
Additionally, traditional long-term risk models may not be as accurate at
estimating a woman’s risk of developing breast cancer, as their average AUC is
around 0.60.[v] ProFound
AI Risk offers a more individualized approach, as it includes a woman’s
mammography images and focuses on one-to-two-years in the future, which
provides critical information that can help clinicians personalize breast
cancer screening regimens for patients based on their individual risk of
developing cancer before or at their next screening.”
Established lifestyle-familial risk models,
such as Tyrer-Cuzick and Gail, are currently used in the U.S. to identify women
with a greater than 20% lifetime risk of developing breast cancer who could be
offered breast magnetic resonance imaging (MRI) as a supplemental screening
modality for breast cancer detection.i However, these long-term risk models can result in a high number of
false positives due to low-to-moderate discrimination performance.[vi]
ProFound AI Risk complements traditional risk models and is easy for clinicians
and medical facilities to adopt, as it only requires the images from a 2D or 3D
mammogram, with no questionnaires, portals, or staff required to implement.
Using U.S. guidelines, the researchers
determined that 14% of the women studied were at high risk after a negative or
benign screening, with an almost 20-fold higher risk of developing breast
cancer, compared to the general risk population. In this high-risk group, 76%
of stage II or later cancers, 59% of stage 0, and 58% of stage 1 cancers were
observed.
Researchers estimated if the 12% of women at
highest risk had been offered supplemental screening based on the ProFound AI
Risk for DBT model, up to 59% of the cancers could have potentially been
detected, compared to 24% of the cancers using Tyrer-Cuzick. This
corresponds to 2.4 times higher sensitivity than Tyrer-Cuzick.
“Our research showed that women
who ProFound AI Risk determined to be at high risk were more likely to present
with later stage tumors than early-stage cancers,” according to lead author of
the study, Mikael Eriksson, PhD, Karolinska Institutet. “It is known that breast
cancer survival is four and 12-fold worse for stage II and III cancers compared
to stage 0 and I cancers in the first four years after diagnosis. Furthermore,
the treatment cost for stage II and III cancers is more than twice that of
stage 0 and I cancers in the first 24 months after diagnosis. ProFound AI Risk
offers the potential to aid radiologists in refining personalizing screening
recommendations and discussing risk with women, which could in turn influence
their screening regimen compliance and
potentially lead to earlier detection, reduced treatment costs and improved
outcomes.”
Researchers also found ProFound AI Risk
provided high accuracy in estimating future risk for invasive and in-situ
cancers, screen detected and interval cancers, and in women with both dense and
non-dense breasts.
“Earlier cancer detection can have a tremendous impact on
women, from treatment to outcomes. And because women are often
caretakers, improving outcomes in women’s health can also have cascading
benefits for children, families, and communities,” added Stevens. “Only iCAD’s Breast
AI suite provides clinicians unprecedented insights into each patient’s present
and future, which offers the potential to transform the trajectory of a woman’s
outcome and life.”
[i] Eriksson, M et al. A
risk model for digital breast tomosynthesis to predict breast cancer and guide
clinical care. Science Translational Medicine. 14 (644). 2022 May 11. Accessed
via https://www.science.org/doi/10.1126/scitranslmed.abn3971.
[ii] Participating U.S. screening
sites: Boca Raton Regional Hospital, Boca Raton, FL; Elizabeth Wende Breast
Care, Rochester, NY; Larchmont, NJ; Zwanger-Pesiri Radiology, Lindenhurst, NY.
[iii] iCAD data on file. Variations
per vendor and population may occur. ProFound AI Risk is a clinical decision
support tool. Information is reviewed by the physician. All care decisions are
up to physicians.
[iv] Breastcancer.org. U.S. Breast Cancer Statistics. Accessed via https://www.breastcancer.org/symptoms/understand_bc/statistics.
[v] Eriksson M, Czene K,
Strand F et al. Identification of Women at High Risk of Breast Cancer Who Need
Supplemental Screening. Radiology. 2020 Sept 8. Accessed via https://doi.org/10.1148/radiol.2020201620.
[vi] M. H. Gail, R. M. Pfeiffer, Breast cancer risk model requirements for
counseling, prevention, and screening. J. Natl. Cancer Inst. 110, 994–1002
(2018).