Premarket evidence — what FDA accepted

Device typesamd

source quote (p.6)

“MammoScreen BD is a software-only device (SaMD) using artificial intelligence to assist radiologists in the interpretation of mammograms.”

Algorithmdeep learning modules trained with very large databases of annotated mammograms

source quote (p.8)

“The system includes 'deep learning' modules for the assessment of the breast tissue composition. These modules are trained with very large databases of annotated mammograms.”

Adaptive (vs locked)No

PCCPYes

source quote (p.1)

“FDA's substantial equivalence determination also included the review and clearance of your Predetermined Change Control Plan (PCCP). Under section 515C(b)(1) of the Act, a new premarket notification is not required for a change to a device cleared under section 510(k) of the Act, if such change is consistent with an established PCCP granted pursuant to section 515C(b)(2) of the Act.”

Cybersecurity addressedNo

Validation studies (3)

Retrospective clinical

n=922 patients

endpoints: Superiority in standalone performance for density assignment of MammoScreen BD compared to a pre-determined reference value (Kappareference = 0.85); Quadratically weighted Cohen's kappa between the density assessment of MammoScreen BD and the established ground truth

standards: IEC 62304:2006/A1:2016- Medical device software - Software life-cycle processes, IEC 62366-1:2015+AMD1:2020- Medical devices - Application of usability engineering to medical devices.

Retrospective clinical

n=500 patients

endpoints: Superiority in standalone performance for density assignment of MammoScreen BD compared to a pre-determined reference value (Kappareference = 0.85); Quadratically weighted Cohen's kappa between the density assessment of MammoScreen BD and the established ground truth

standards: IEC 62304:2006/A1:2016- Medical device software - Software life-cycle processes, IEC 62366-1:2015+AMD1:2020- Medical devices - Application of usability engineering to medical devices.

Retrospective clinical

n=376 patients

endpoints: Superiority in standalone performance for density assignment of MammoScreen BD compared to a pre-determined reference value (Kappareference = 0.85); Quadratically weighted Cohen's kappa between the density assessment of MammoScreen BD and the established ground truth

standards: IEC 62304:2006/A1:2016- Medical device software - Software life-cycle processes, IEC 62366-1:2015+AMD1:2020- Medical devices - Application of usability engineering to medical devices.

Reported performance (9 observations)

accuracyas written: “Accuracy (four-class)”85.4CI [82.00, 88.40]

source quote (p.10)

“Accuracy = 85.40 with CI = [82.00, 88.40]”

agreement_kappaas written: “Cohen's Kappa (quadratic, four-class)”89.54CI [86.88, 91.69]

source quote (p.10)

“Cohen's Kappa (quadratic) = 89.54 with CI = [86.88, 91.69]”

agreement_kappaas written: “Cohen's Kappa (linear, four-class)”83.75CI [79.85, 87.00]

source quote (p.10)

“Cohen's Kappa (linear) = 83.75 with CI = [79.85, 87.00]”

accuracyas written: “Accuracy (binary)”93CI [90.80, 95.20]

source quote (p.10)

“Accuracy = 93.00 with CI = [90.80, 95.20]”

agreement_kappaas written: “Cohen's Kappa (quadratic, binary)”86CI [81.36, 90.25]

source quote (p.10)

“Cohen's Kappa (quadratic) = 86.00 with CI = [81.36, 90.25]”

agreement_kappaas written: “Cohen's Kappa (linear, binary)”86CI [81.36, 90.25]

source quote (p.10)

“Cohen's Kappa (linear) = 86.00 with CI = [81.36, 90.25]”

agreement_kappaas written: “Quadratically weighted Cohen's kappa (Hologic)”89.03CI [87.43 – 90.56]

source quote (p.11)

“Kappa quadratic = 89.03 [95% CI: 87.43 – 90.56]”

agreement_kappaas written: “Quadratically weighted Cohen's kappa (Hologic Envision)”89.54CI [86.88–91.69]

source quote (p.11)

“Kappa quadratic = 89.54 [95% CI: 86.88–91.69]”

agreement_kappaas written: “Quadratically weighted Cohen's kappa (GE)”93.19CI [90.50–94.92]

source quote (p.11)

“Kappa quadratic = 93.19 [95% CI: 90.50–94.92]”

Each value carries its own analysis unit and task — never compare or pool across devices. Source: 510(k) summary PDF.

Predicate network

Postmarket — what happened after clearance

0

recalls in product code, 24mo

3

MAUDE reports in code, 12mo

—

vs code's own 3-yr baseline

0

drift signals on this device

Recall and MAUDE counts are product-code-level (reports aren't reliably attributable to one device). Signals are descriptive observables with sources — never a judgment that the device is unsafe or drifting. Snapshot 2026-07-08.

Reimbursement — how devices like this got paid

Not yet tracked — no payment pathway indexed for this clearance (the reimbursement corpus is a growing seed set).