BunkerHill BMD

K242295

BunkerHill Health · cleared 2025-04-08 · product code KGI · Radiology

Premarket evidence — what FDA accepted

Device typesamd
source quote (p.5)
“The Bunkerhill BMD application is a software only medical device (SaMD) that includes deep- learning-based computer vision and post-processing algorithms that estimates the bone mineral density from previously obtained computed tomography (CT) images.”
Algorithmdeep- learning-based computer vision and post-processing algorithms
source quote (p.5)
“The Bunkerhill BMD application is a software only medical device (SaMD) that includes deep- learning-based computer vision and post-processing algorithms that estimates the bone mineral density from previously obtained computed tomography (CT) images.”
Adaptive (vs locked)FDA source did not state this
PCCPFDA source did not state this
Cybersecurity addressedYes
source quote (p.8)
“Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions.”

Validation studies (1)

Retrospective clinical

n=371 cases · 4 site(s)

endpoints: Sensitivity; Specificity; Pearson correlation coefficient; AUROC; PPV; NPV

standards: Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices, Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions

Reported performance (5 observations)

sensitivity81CI 74.0 - 86.8
source quote (p.8)
“The Bunkerhill BMD algorithm achieved a sensitivity of 81.0 (74.0 - 86.8) and specificity of 78.4 (72.3 - 83.7), which passed the acceptance criteria for the primary endpoint with lower bound 95% confidence interval of both Sensitivity and Specificity being greater than 70%.”
specificity78.4CI 72.3 - 83.7
source quote (p.8)
“The Bunkerhill BMD algorithm achieved a sensitivity of 81.0 (74.0 - 86.8) and specificity of 78.4 (72.3 - 83.7), which passed the acceptance criteria for the primary endpoint with lower bound 95% confidence interval of both Sensitivity and Specificity being greater than 70%.”
aurocas written: “auc”0.883CI 0.849–0.916
source quote (p.8)
“Additionally, the device achieved was evaluated across multiple secondary metrics, including a Pearson correlation coefficient of 0.791 (95% CI: 0.752–0.830), AUROC of 0.883 (95% CI: 0.849–0.916), PPV of 73.6% (95% CI: 66.4%–79.9%), and NPV of 84.8% (95% CI: 79.0%–89.5%), further supporting the robustness and reliability of the algorithm.”
ppvas written: “PPV”73.6CI 66.4%–79.9%
source quote (p.8)
“Additionally, the device achieved was evaluated across multiple secondary metrics, including a Pearson correlation coefficient of 0.791 (95% CI: 0.752–0.830), AUROC of 0.883 (95% CI: 0.849–0.916), PPV of 73.6% (95% CI: 66.4%–79.9%), and NPV of 84.8% (95% CI: 79.0%–89.5%), further supporting the robustness and reliability of the algorithm.”
npvas written: “NPV”84.8CI 79.0%–89.5%
source quote (p.8)
“Additionally, the device achieved was evaluated across multiple secondary metrics, including a Pearson correlation coefficient of 0.791 (95% CI: 0.752–0.830), AUROC of 0.883 (95% CI: 0.849–0.916), PPV of 73.6% (95% CI: 66.4%–79.9%), and NPV of 84.8% (95% CI: 79.0%–89.5%), further supporting the robustness and reliability of the algorithm.”

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

1
recalls in product code, 24mo
6
MAUDE reports in code, 12mo
+200%
vs code's own 3-yr baseline
1
drift signals on this device
  • recall_reason_pattern

    Software/algorithm-related recall in product code KGI (Medimaps Group Fongit Chemin des Aulx 18 Plan-les-Ouates Switzerland, initiated 2025-02-03): "Potential variability in calculations from fast array scans compared to array scans when operating on Hologic Horizon machines." Recalling firm is another firm in the same product code.

    first seen 2026-07-08 · recall res_event_number:96233

Recall and MAUDE counts are product-code-level (reports aren't reliably attributable to one device); a recall is shown as device-attributed only when the recall record itself lists this clearance number. 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).

Applicable FDA guidance — what the submission is measured against

FDA guidance documents and guiding principles applicable to 510(k) AI/ML devices in the Radiology panel. A curated reference index, not legal or regulatory advice — each item states its own status, and a draft is never binding.

Applicability is derived from the device's FDA advisory panel and pathway — cross-cutting guidances apply to every AI/ML device; panel-specific ones are flagged. Titles, dates, and links verified against fda.gov as of July 2026.

Constat Precedent · public FDA/CMS data · descriptive decision-support, not regulatory or reimbursement advice. Share this page: constat.dev/precedent/device/K242295