QAS: Radiological computer-assisted triage and notification software
What it covers
Software that analyzes medical images, such as CT angiograms of the head or chest CT, and notifies a specialist of a suspected time-sensitive finding so the case can be prioritized. It runs in parallel to standard of care and does not provide a diagnosis.
Recent clearances
| 510(k) | Device | Applicant | Decision | Type |
|---|---|---|---|---|
| K263109 | EFAI NEUROSUITE CTA LVO ASSESSMENT SYSTEM (LVO-CTA-100) | Ever Fortune.Ai, Co., Ltd. | Oct 30, 2026 | Traditional |
| K262301 | Rapid ICH (v4.0) | Ischemaview, Inc. | Oct 9, 2026 | Traditional |
| K260906 | a2z-Abdo-Triage | A2z Radiology Ai, Inc. | Sep 16, 2026 | Traditional |
| K261085 | Rapid Vessel Occlusion (Rapid VO) | iSchemaview | Jul 13, 2026 | Traditional |
| K260729 | Vascular Assist Occlusion Triage (VAOT) (v1.0) | GuideAI Health | Jun 18, 2026 | Traditional |
| K261317 | BriefCase-Triage | Aidoc Medical , Ltd. | May 14, 2026 | Special |
| K253163 | EFAI ERSUITE CT APPENDICITIS ASSESSMENT SYSTEM (APPEN-CT-100) | Ever Fortune.Ai, Co., Ltd. | Apr 22, 2026 | Traditional |
| K252421 | JLK-NCCT | JLK, Inc. | Mar 24, 2026 | Traditional |
| K253818 | Annalise Enterprise | Harrison-AI Medical Pty, Ltd. | Mar 3, 2026 | Traditional |
| K253578 | BriefCase-Triage: CARE Multi-Triage CT for Pneumothorax; Pericardial effusion; Large aortic aneurysm; Shoulder fracture or dislocation device | Aidoc Medical , Ltd. | Feb 26, 2026 | Traditional |
Clearances by year
89 QAS 510(k)s were cleared from 2011 to Oct 29, 2026. Days to decision run from the day FDA received the 510(k) to its decision, including time on hold.
| Year | Clearances | Median days to decision |
|---|---|---|
| 2026 (so far) | 11 | 104 |
| 2025 | 15 | 101 |
| 2024 | 12 | 136 |
| 2023 | 11 | 98 |
| 2022 | 15 | 116 |
| 2021 | 5 | 86 |
| 2020 | 12 | 112 |
| 2019 | 5 | 111 |
| 2018 | 3 | 137 |
Who submits
32 applicants cleared a QAS device since 2011. 31% of the clearances went to applicants in Israel, then United States (28%), South Korea (7%) and Taiwan (7%). By type: Traditional 83%, Special 16% and Direct 1%.
| Applicant | Clearances since 2011 |
|---|---|
| Aidoc Medical , Ltd. | 25 |
| Ischemaview, Inc. | 8 |
| Avicenna.Ai | 5 |
| Ever Fortune.Ai, Co., Ltd. | 5 |
| Annalise-Ai Pty , Ltd. | 4 |
| Brainomix Limited | 4 |
| JLK, Inc. | 4 |
| Viz. Ai, Inc. | 4 |
Testing FDA usually expects
- Standalone performance on a multi-site retrospective dataset kept separate from training data, reported as sensitivity and specificity with confidence intervals
- Time to notification compared with standard of care
- Subgroup analysis by age, sex, scanner manufacturer and acquisition protocol
- Software verification and validation per IEC 62304, and cybersecurity documentation for connected software
Standards
QAS is one of the product codes most common on FDA’s list of AI-enabled devices. See how FDA regulates AI-enabled medical devices.
Ask about QAS
Questions about QAS
What is FDA product code QAS?
Product code QAS covers radiological triage and notification software, Class II devices regulated under 21 CFR 892.2080 and reviewed by FDA's Radiology panel. FDA's 510(k) database lists 90 clearances under QAS.
How many 510(k) clearances are there under QAS?
FDA’s 510(k) database listed 90 clearances under QAS as of Nov 16, 2026, 89 of them decided since 2011; the most recent decision was Oct 30, 2026. The Predicate Finder searches the live database.
How long does FDA take to clear a QAS 510(k)?
For the 28 QAS clearances decided in the two years before Oct 29, 2026, the median time from FDA receiving the 510(k) to its decision was 104 days. That counts calendar days, including time the submission was on hold waiting for answers from the submitter.
Which companies have the most QAS clearances?
Since 2011: Aidoc Medical , Ltd. (25), Ischemaview, Inc. (8), Avicenna.Ai (5), Ever Fortune.Ai, Co., Ltd. (5) and Annalise-Ai Pty , Ltd. (4). 32 applicants have cleared a QAS device in that time.
What testing do QAS submissions include?
Recent QAS 510(k) summaries typically describe: standalone performance on a multi-site retrospective dataset kept separate from training data, reported as sensitivity and specificity with confidence intervals; time to notification compared with standard of care; subgroup analysis by age, sex, scanner manufacturer and acquisition protocol; software verification and validation per IEC 62304, and cybersecurity documentation for connected software. Check the summaries of your own predicates for what they submitted.