HeartFocus (V.1.1.1)

K242807

Deski · cleared 2025-04-04 · product code QJU · Radiology

Premarket evidence — what FDA accepted

Device typesamd
source quote (p.5)
The HeartFocus software is a radiological computer-assisted acquisition guidance system that provides real-time user guidance during echocardiography to assist the user in acquiring anatomically standard diagnostic-quality 2D echocardiographic views. HeartFocus software is an accessory to compatible general-purpose diagnostic ultrasound systems.
AlgorithmArtificial intelligence (AI) and machine learning models, including deep learning, to emulate sonographer expertise for probe positioning, diagnostic-quality view detection, and recording.
source quote (p.6)
HeartFocus uses artificial intelligence (AI) to emulate the expertise of sonographers in positioning the probe on the patient's chest and in identifying and recording diagnostic-quality clips.
Adaptive (vs locked)No
source quote (p.21)
M1 - Retraining of the core algorithms: This modification's scope is the retraining of the AI/ML models in the perspective of validating a new manufacturer prior to an M2. The modification is limited to retraining with additional data without changing the models' architecture or training procedure.
PCCPYes
source quote (p.1)
FDA's substantial equivalence determination also included the review and clearance of your Predetermined Change Control Plan (PCCP).
Cybersecurity addressedYes
source quote (p.6)
HeartFocus is an application that operates entirely offline, without requiring a cloud server to provide its functionalities. All collected medical data is stored locally on the tablet. This data is never transferred to a server controlled by DESKi.

Validation studies (3)

Bench

n=31 other

endpoints: Success rate of critical tasks

standards: FDA Guidance Document, « Applying Human Factors and Usability Engineering to Medical Devices », IEC 62366 1:205 standard

Standalone

n=290 patients

endpoints: Cohen's kappa score; Positive predictive value of successful guidance cues; Positive predictive value of high-quality records

Prospective clinical

n=240 patients · 2 site(s)

endpoints: Qualitative Visual Assessment of LV Size; Qualitative Visual Assessment of LV Function; Qualitative Visual Assessment of RV Size; Qualitative Visual Assessment of Non-Trivial Pericardial Effusion

Reported performance (5 observations)

agreement_kappaas written: “Cohen's kappa score for Diagnostic-quality view detectionstated without valueCI 0.699 [0.673, 0.724] to 0.873 [0.861, 0.884]
source quote (p.15)
Cohen's kappa scores range from 0.699 [0.673, 0.724] to 0.873 [0.861, 0.884], meeting the success criteria of Cohen's kappa score > 0.6 on the lower bound of the 95% CI for each reference view.
ppvas written: “Positive predictive value (PPV) for Live guidance cuesstated without valueCI 0.810 [0.804, 0.816] to 0.953 [0.951, 0.955]
source quote (p.15)
Guidance cues PPV ranges from 0.810 [0.804, 0.816] to 0.953 [0.951, 0.955], satisfying the success criteria of PPV > 0.8 on the lower bound of the 95% CI for each reference view.
ppvas written: “Positive predictive value (PPV) for Auto record featurestated without valueCI 0.846 [0.665, 0.938] to 1.000 [0.908, 1.000]
source quote (p.15)
While using the Auto record feature solely, the PPV ranges from 0.846 [0.665, 0.938] to 1.000 [0.908, 1.000], meeting the success criteria of PPV > 0.6 on the lower bound of the 95% CI and PPV > 0.8 on the point estimate for each reference view.
ppvas written: “Positive predictive value (PPV) for Auto record and Best-Effort record combinedstated without valueCI 0.816 [0.666, 0.908] to 1.000 [0.914, 1.000]
source quote (p.15)
While using both the Auto record and Best-Effort record, the PPV ranges from 0.816 [0.666, 0.908] to 1.000 [0.914, 1.000].
specificityas written: “Specificity of Auto record feature (diagnostic quality clips)99.7
source quote (p.20)
99.7% of the clips recorded with the “Auto record” feature were of diagnostic quality, demonstrating the high specificity of this feature.

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
0
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).

RIGOR™ Precedent · public FDA/CMS data · descriptive decision-support, not regulatory or reimbursement advice. Share this page: radar.healthai.com/precedent/device/K242807