ScreenPoint Medical’s research roots run deep: as a spin-off of Radboud University Medical Center, the company consistently strives to build products that are not only grounded in real-world problems but substantiated by rigorous scientific scrutiny.
As this body of evidence has grown significantly over the past year and with the introduction of Transpara 2.1 (which offers notable performance increases in sensitivity for both 2D and 3D mammography compared to 1.7.4), we are revisiting key peer-reviewed studies from 2025 that grew our understanding of the impact of Transpara Breast AI on cancer detection and radiology workflow.
Early detection of breast cancers between screenings – data from UCLA
In a retrospective analysis, Yu (2025) finds that within a cancer-enriched dataset, Transpara 1.7.1 flagged 76% of 3D mammograms that had been originally read as normal but were later linked to an interval breast cancer.
It also flagged 90% of missed reading error cases where the cancer had been visible on the mammogram but missed or misinterpreted by the radiologist. Incorporating Transpara into screening could help reduce the number of mammographically visible types of breast cancer by up to 30%.
Learn more or view Imaging Wire's interview with UCLA's Dr. Hannah Milch.
Reliable identification of negative cases – data from Athena Breast Health Network
Based on a population-based DBT screening cohort study of 4,824 exams from the Athena Breast Health Network (California), Chen (2025) reports that Transpara 1.7.1 attained a sensitivity of 89.2%, specificity of 68.5%, and negative predictive value (NPV) of 99.8% for cases assessed by Transpara as Intermediate or Elevated Risk. Authors highlight that Transpara Detection “could reliably identify negative DM and DBT examinations, potentially aiding radiologists’ workflow efficiency” but also highlight the need for strategies to manage patient recall associated with Intermediate Risk cases. Learn more.
High-risk case triage – data from Mount Sinai
Most recently, Mathur (2025) outlines results from the Icahn School of Medicine at Mount Sinai using Transpara 1.7.0 on a high-risk subset of 3D exams: “This real-world retrospective study evaluated the standalone performance of Transpara AI in detecting breast cancer on screening digital breast tomosynthesis (DBT) exams which proceeded to stereotactic biopsy. Within this high-risk subset, Transpara AI had high sensitivity and high negative predictive value of approximately 95%.” Learn more.
The most rigorous analysis attests to Transpara’s performance – MASAI Trial
In the landmark MASAI study from Sweden, Transpara 1.7.0 became the first Breast AI solution validated and vetted by a randomized controlled trial. A trial update from Hernström (2025) finds that Transpara Detection “contributes to the early detection of clinically relevant breast cancer and reduces screen-reading workload without increasing false positives” with a 29% increase in cancer detection and a 44% workload reduction compared to standard double reading across a population of over 100,000 women. Learn more.
Comparable performance to double reading – data from The Netherlands
Across a Dutch cohort of over 40,000 FFDM exams using Transpara 1.7.0, research published in The Lancet Digital Health shows that "AI detection of breast cancer in population-based mammography screening is comparable with double human reading. AI misses some breast cancers that are recalled by human-assessment but detects a similar number of breast cancers otherwise missed by the interpreting radiologists. This is in line with similar (double-read) screening studies also reporting that AI support improves detection rates and reduces false negatives." Learn more.
Transpara is cost-effective and clinically impactful – data from Sweden
A first-of-its-kind, simulated economic analysis of the cost-effectiveness of Breast AI in screening finds that Transpara 1.7.0 could drive cost savings of €59,320 per 1,000 women screened and lead to an increase of 10.8 quality adjusted life years per 1,000 women. Break-even on cost occurred at the second screening (42 years of age), suggesting that AI adoption is both a dominant strategy and a “cost-saving strategy compared to a conventional strategy using double human screen reading.” Learn more.
Risk-based screening for supplemental MRI – data from The Netherlands
In Radiology, van Winkel (2025) suggests that using Transpara 1.7.0 “to select women for supplemental MRI effectively identified women with higher breast cancer risk in an intermediate-risk population, including women with mammographically occult cancers. AI selection of women with intermediate risk for supplemental MRI screening has the potential to reduce screening burden and costs, while maintaining a high cancer detection rate.” Learn more.
DCIS Invasion Rule-out – data from Turkey
In a proof-of-concept study, retrospective study, Gundogdu (2025) finds a high negative predictive value of 96.2% when using Transpara 1.7.4 as a rule-out test for invasion in ductal carcinoma in situ (DCIS) cases. Study authors note that “With this high NPV, clinicians could potentially identify patients with a very low probability of invasive disease who may be suitable candidates for less aggressive treatment approaches or active surveillance strategies, thus supporting treatment de-escalation decisions in appropriate clinical contexts. This approach has the potential to reduce morbidity associated with unnecessary treatments and enhance the effective utilization of healthcare resources.” Learn more.
In 2025, studies touched on a variety of topics, including:
For an even deeper dive, ScreenPoint Medical’s VP of Clinical Strategy offers a tour-de-force of Transpara’s clinical evidence over the past several years in an on-demand webinar recorded in October 2025.
Watch now or visit our Evidence Page for a full listing of publications.