JLK-NCCT

Authorization

K252421 added LVO notification to an existing ICH workflow; its 288-case evaluation did not revalidate the unchanged ICH model.

Submission lineageK252421 includes the unchanged ICH component cleared in K243363.

Intended use

Analyze adult non-contrast head CT in parallel with standard care and notify clinicians of suspected intracranial hemorrhage or ICA, M1 or M2 large-vessel occlusion.

Input

Adult non-contrast head CT DICOM series

Output

Mobile suspected-ICH notification; Mobile suspected-LVO notification; Compressed non-diagnostic preview images

Use boundary

Notification only: a negative alert does not exclude ICH or LVO, the preview is non-diagnostic, and CTA or MRA remains required in the standard ischemic-stroke workup.

AI disclosure and function

AI/ML explicit

The summary explicitly describes AI/ML models for ICH, hypodensity and hyperdense-artery-sign analysis in a cascade notification workflow.

The AI output is a triage flag and non-diagnostic preview, not a diagnostic report, treatment recommendation or generative interface.

No foundation-model, LLM, or generative-AI function was identified in the reviewed FDA materials.

Validation

Independent retrospective standalone and reader-performance evaluation for the LVO component; the ICH component was inherited unchanged from K243363.

Comparator

Consensus ground truth and performance of neuroradiologist and general-radiologist readers; notification time was compared with a predicate performance goal.

Ground truth

Two US board-certified neuroradiologists with a third adjudicator for disagreements.

Representation

The public summary identifies an independent 288-case LVO dataset but does not provide a complete demographic table in the reviewed section.

Reported performance

LVO standalone sensitivity

78.5% (95% CI 71.9%-84.7%)Retrospective detection performance on 144 LVO-positive and 144 LVO-negative cases.

LVO standalone specificity

90.3% (95% CI 85.1%-94.7%)Retrospective detection performance, not exclusion of disease after a negative notification.

LVO standalone AUC

0.880 (95% CI 0.837-0.920)Discrimination on the independent LVO test set.

NCCT-to-notification time

1.67 +/- 0.10 minutesTechnical processing and delivery time, not treatment-time or outcome improvement.

Limitations

• The reported 288-case clinical evaluation addressed LVO, not a new validation of the unchanged ICH algorithm.
• The system is notification only and cannot rule out ICH or LVO.
• No prospective workflow, treatment-time or neurological-outcome study was identified in the clearance summary.

It does not prove autonomous diagnosis, exclusion of disease, faster reperfusion or improved neurological outcomes.

PCCP / Change control

No authorized PCCP was identified for this submission in the reviewed FDA database record or decision material.

Absence of an identified PCCP does not prevent every future device change; it means no authorized plan was established in the reviewed public record.

Postmarket check

No exact-name openFDA device-enforcement result was found in the targeted query.

No exact-name openFDA result found as of 2026-09-08; this is not proof that no recall exists.

Primary FDA sources

1. FDA AI-enabled device list — Submission K252421Confirms inclusion, list display name, company, lead panel, product code and decision date only.
2. FDA submission database — 510(k) K252421Confirms pathway, review panel, decision and PCCP database field.
3. FDA decision summary — 510(k) summary pp. 1-5; limitations pp. 1-3; performance pp. 4-5Supports intended use, AI workflow, notification-only limits, component provenance, design, performance and reader comparison.

Sources checked Sep 8, 2026 · Next scheduled review Oct 8, 2026. “No exact-name result” is a bounded search finding, not proof that no safety event or recall exists.