MammoScreen

Performance Metrics

Clinical Evidence

The published evidence is anchored by a single peer-reviewed study — the manufacturer-authored multireader, multicase evaluation behind the original clearance (Pacilè et al., Radiology: Artificial Intelligence 2020). Fourteen radiologists read an enriched set of 240 digital mammography exams with and without the software in a counterbalanced design: mean reader AUC rose from 0.769 unassisted to 0.797 assisted, a difference of 0.028 (95% CI 0.002 to 0.055, P = .035). It is a competent MRMC study in a serious venue, and it is also everything a pivotal reader study usually is: enriched, retrospective, and funded by the manufacturer. The tomosynthesis clearance (K211541, November 2021) rests on a second multireader study that was reviewed by the FDA but, as far as could be established this run, never journal-published — the DBT performance record lives in the 510(k) summary, conference posters and press material. The remainder of the manufacturer's research output is conference-grade, including an IWBI 2022 proceedings paper on using temporal change information from prior exams to cut false positives — the capability that became the priors feature of the current generation. What is missing is independent evidence. An MGH-group evidence review of FDA-cleared DBT CADe/x tools (Lamb et al., 2024) situates MammoScreen among the six cleared systems but reports no independent performance study of it, and none was located in this review. That is the honest gap against the catalogue's other mammography entries: ProFound AI carries independent head-to-head comparisons, and Transpara carries the 105,000-woman MASAI randomised trial. MammoScreen's four-generation regulatory cadence has so far outpaced its evidence base — the inverse of the pattern its rivals set, and the reason its mark sits where it does.

Clinical Pulse

Verified clinician sentiment

Verified clinician reviews launching soon · Apply to contribute →

Reviews are signed by clinicians verified against their regulator's public register.

Inside the Auris+ Listing

Regulatory Approvals

Safety Record

No MammoScreen recall, FDA safety communication or adverse-event signal was found in indexed public sources as of July 2026 — read as "none found in public reporting" rather than an exhaustive audit, since the MAUDE and FDA enforcement databases could not be queried directly from this environment. As a concurrent-read aid the device does not remove exams from the worklist, suppress findings or alter source images, which limits its harm surface. The residual risks are the class-standard ones: a low MammoScreen Score is not a normal-anatomy finding, and over-reliance on it risks premature dismissal of subtle cancers — the cleared labelling's instruction that management decisions must not rest solely on the software is the operative safeguard.

Intended Use & Indications

MammoScreen is a software-only device used by interpreting physicians as a concurrent reading aid for screening full-field digital mammography and digital breast tomosynthesis exams acquired on compatible systems. Per the FDA 510(k) record, it identifies findings suspicious for breast cancer — soft-tissue lesions and calcification groups — places marks on them, and assesses their level of suspicion, expressed per finding, per breast and for the exam overall. The cleared use is assistive throughout: the interpreting physician reads every exam, and patient-management decisions must not be made solely on the basis of MammoScreen's analysis. Unlike several catalogue-mates whose US clearances are triage-only, MammoScreen's US scope is the full concurrent-reading aid; it has no autonomous or worklist-prioritisation clearance, and its workflow features (complexity index, pre-reporting) sit around the read rather than replacing any part of it.