MammoScreen BD
K241561Therapixel · cleared 2024-10-02 · product code QIH · Radiology
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.”
AlgorithmImage feature-based with deep learning
source quote (p.8)
“Image feature-based with deep learning”
Adaptive (vs locked)No
source quote (p.8)
“Predetermined Change Control Plan (PCCP) including: • Support of GE mammograms (no re-training required) • Support of Siemens mammograms (re-training required) • Pre-training of backbone using Unsupervised Machine Learning (as opposed to Supervised Machine Learning)”
PCCPYes
source quote (p.21)
“MammoScreen BD is powered by machine-learning neural architectures. Therapixel will make future algorithm improvements under a PCCP. The plan describes the future modifications, assesses their impact, and a modification protocol details how data management, re-training, performance evaluation and update procedures will be handled.”
Cybersecurity addressedFDA source did not state this
Validation studies (1)
Retrospective clinical
n=922 patients · 3 site(s)
endpoints: accuracy; reproducibility; agreement with ground truth; confusion matrix; correctly classified examinations
standards: IEC 62304:2006/A1:2016
Reported performance (21 observations)
agreement_kappaas written: “quadratically-weighted Cohen's kappa (four-class)”0.8903CI [87.43 – 90.56]
source quote (p.10)
“On the four-class task, MammoScreen BD achieved a quadratically-weighted Cohen's kappa of 89.03, 95% confidence interval [87.43 – 90.56].”
accuracyas written: “Accuracy (four-class)”0.8468CI [82.68, 86.67]
source quote (p.11)
“Accuracy = 84.68 with CI = [82.68, 86.67]”
agreement_kappaas written: “Cohen's Kappa (quadratic, four-class)”0.8903CI [87.43, 90.56]
source quote (p.11)
“Cohen's Kappa (quadratic) = 89.03 with CI = [87.43, 90.56]”
agreement_kappaas written: “Cohen's Kappa (linear, four-class)”0.8294CI [80.67, 85.22]
source quote (p.11)
“Cohen's Kappa (linear) = 82.94 with CI = [80.67, 85.22]”
agreement_kappaas written: “quadratically-weighted Cohen's kappa (binary)”0.845CI [81.46, 87.36]
source quote (p.11)
“On the binary classification task, MammoScreen BD achieved a quadratically-weighted Cohen's kappa of 84.50, 95% confidence interval [81.46, 87.36].”
accuracyas written: “Accuracy (binary)”0.9229CI [90.82, 93.77]
source quote (p.11)
“Accuracy = 92.29 with CI = [90.82, 93.77]”
agreement_kappaas written: “Cohen's Kappa (linear, binary)”0.845CI [81.46, 87.36]
source quote (p.11)
“Cohen's Kappa (linear) = 84.50 with CI = [81.46, 87.36]”
accuracyas written: “Accuracy (FFDM only)”0.8485CI [82.09, 87.33]
source quote (p.13)
“Accuracy = 84.85 with CI = [82.09, 87.33]”
agreement_kappaas written: “Cohen's Kappa (quadratic, FFDM only)”0.8911CI [87.01, 91.00]
source quote (p.13)
“Cohen's Kappa (quadratic) = 89.11 with CI = [87.01, 91.00]”
agreement_kappaas written: “Cohen's Kappa (linear, FFDM only)”0.8308CI [79.99, 85.84]
source quote (p.13)
“Cohen's Kappa (linear) = 83.08 with CI = [79.99, 85.84]”
agreement_kappaas written: “Cohen's Kappa (quadratic, 2DSM only)”0.8889CI [85.88 – 91.50]
source quote (p.13)
“The same metric achieves a value of 88.89, 95% CI [85.88 – 91.50] when evaluated on 2DSM mammograms (see the figure below).”
agreement_kappaas written: “Cohen's Kappa (quadratic, Age < 55)”0.8793CI [85.19 – 90.26]
source quote (p.14)
“Age <55: 87.93 (95% CI: 85.19 – 90.26)”
agreement_kappaas written: “Cohen's Kappa (quadratic, 55 <= Age < 65)”0.9122CI [88.05 – 93.77]
source quote (p.14)
“55 ≤ Age < 65: 91.22 (95% CI: 88.05 – 93.77)”
agreement_kappaas written: “Cohen's Kappa (quadratic, Age >= 65)”0.8715CI [82.70 – 90.96]
source quote (p.14)
“Age ≥ 65: 87.15 (95% CI: 82.70 – 90.96)”
agreement_kappaas written: “Cohen's Kappa (quadratic, Thickness < 50 mm)”0.8581CI [81.99 – 89.30]
source quote (p.16)
“Thickness < 50 mm: 85.81 (95% CI: 81.99 – 89.30)”
agreement_kappaas written: “Cohen's Kappa (quadratic, 50 mm <= Thickness < 70 mm)”0.869CI [84.25 – 89.34]
source quote (p.16)
“50 mm ≤ Thickness < 70 mm: 86.90 (95% CI: 84.25 – 89.34)”
agreement_kappaas written: “Cohen's Kappa (quadratic, Thickness >= 70 mm)”0.8929CI [85.19 – 92.89]
source quote (p.16)
“Thickness ≥ 70 mm: 89.29 (95% CI: 85.19 – 92.89)”
agreement_kappaas written: “Cohen's Kappa (quadratic, White race)”0.8788CI [81.82 – 93.18]
source quote (p.18)
“White: 87.88 (95% CI: 81.82 – 93.18)”
agreement_kappaas written: “Cohen's Kappa (quadratic, Asian race)”0.8711CI [80.22 – 93.04]
source quote (p.18)
“Asian: 87.11 (95% CI: 80.22 – 93.04)”
agreement_kappaas written: “Cohen's Kappa (quadratic, Black or African American race)”0.9082CI [85.58 – 94.97]
source quote (p.18)
“Black or African American: 90.82 (95% CI: 85.58 – 94.97)”
agreement_kappaas written: “Cohen's Kappa (quadratic, American Indian or Alaska native or native Hawaiian or Pacific islander race)”0.5385CI [0.00 - 83.33]
source quote (p.18)
“American Indian or Alaska native or native Hawaiian or Pacific islander: 53.85 (95% CI: 0.00 - 83.33)”
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
2
drift signals on this device
- re_clearance
The FDA AI/ML device list shows a newer 510(k) K243685 (decision 2025-08-22) from Therapixel for a matching device line ("MammoScreen BD") — a new clearance for the same line is a change event.
first seen 2026-07-08 · k_number:K243685
- re_clearance
The FDA AI/ML device list shows a newer 510(k) K243679 (decision 2025-07-03) from Therapixel for a matching device line ("MammoScreen® (4)") — a new clearance for the same line is a change event.
first seen 2026-07-08 · k_number:K243679
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).