Multi-national (US and EU), multicenter clinical validation study: 244 patients with 311 target vessels from 8 clinical sites.
—
Indications for Use
DEEPVESSEL FFR is a coronary physiological simulation software for the clinical quantitative and qualitative analysis of previously acquired Computed Tomography DICOM data for clinically stable symptomatic patients with coronary artery disease. It provides DVFFR (a CT-derived FFR measurement) computed from static coronary CTA images using deep learning neural networks that encode imaging, structural, and functional characteristics of coronary arteries through learning. DEEPVESSEL FFR analysis is intended to support the functional evaluation of coronary artery disease. The results of the analysis are provided to support qualified clinicians to aid in the evaluation and assessment of coronary arteries. DEEPVESSEL FFR results are intended to be used by qualified clinicians in conjunction with the patient's clinical history, symptoms, and other diagnostic tests, as well as the clinician's professional judgment.
Device Story
DEEPVESSEL FFR is a coronary physiological simulation software that processes previously acquired coronary CTA DICOM images. It uses deep learning neural networks to automatically generate a 3D coronary artery tree model; qualified image analysts may manually edit segmentations. The system then estimates semi-continuous FFR values (DVFFR) along coronary centerlines using a multi-layer perceptron (MLP) and a bidirectional multi-layer recursive neural network (BRNN). Outputs include a 3D DVFFR tree model and a PDF report sent via a third-party service portal. Used by clinicians to aid in the functional evaluation of coronary artery disease in conjunction with clinical history and other diagnostic tests. The device provides a non-invasive method to assess coronary artery function, potentially benefiting patients by supporting clinical decision-making.
Clinical Evidence
Multi-national, multicenter clinical validation study (244 patients, 311 vessels). Primary endpoints: per-vessel sensitivity and specificity vs. invasive FFR. Results: per-vessel sensitivity 86.9% (95% CI: 80.6%-92.7%), specificity 86.7% (95% CI: 82.0%-91.1%). Patient-level sensitivity 87.4%, specificity 83.7%. All performance targets met.
Technological Characteristics
Software-based medical device. Uses deep learning neural networks (MLP and BRNN) for 3D coronary artery tree segmentation and FFR estimation. Independent of imaging hardware/vendors. Output: 3D model and PDF report. Connectivity: third-party service portal.
Indications for Use
Indicated for clinically stable symptomatic patients with coronary artery disease to support functional evaluation via CT-derived FFR measurement.
Regulatory Classification
Identification
A coronary vascular physiologic simulation software device is a prescription device that provides simulated functional assessment of blood flow in the coronary vascular system using data extracted from medical device imaging to solve algorithms and yield simulated metrics of physiologic information (e.g., blood flow, coronary flow reserve, fractional flow reserve, myocardial perfusion). A coronary vascular physiologic simulation software device is intended to generate results for use and review by a qualified clinician.
Special Controls
*Classification.* Class II (special controls). The special controls for this device are:(1) Adequate software verification and validation based on comprehensive hazard analysis, with identification of appropriate mitigations, must be performed, including:
(i) Full characterization of the technical parameters of the software, including:
(A) Any proprietary algorithm(s) used to model the vascular anatomy; and
(B) Adequate description of the expected impact of all applicable image acquisition hardware features and characteristics on performance and any associated minimum specifications;
(ii) Adequate consideration of privacy and security issues in the system design; and
(iii) Adequate mitigation of the impact of failure of any subsystem components (
*e.g.,* signal detection and analysis, data storage, system communications and cybersecurity) with respect to incorrect patient reports and operator failures.(2) Adequate non-clinical performance testing must be provided to demonstrate the validity of computational modeling methods for flow measurement; and
(3) Clinical data supporting the proposed intended use must be provided, including the following:
(i) Output measure(s) must be compared to a clinically acceptable method and must adequately represent the simulated measure(s) the device provides in an accurate and reproducible manner;
(ii) Clinical utility of the device measurement accuracy must be demonstrated by comparison to that of other available diagnostic tests (
*e.g.,* from literature analysis);(iii) Statistical performance of the device within clinical risk strata (
*e.g.,* age, relevant comorbidities, disease stability) must be reported;(iv) The dataset must be adequately representative of the intended use population for the device (
*e.g.,* patients, range of vessel sizes, imaging device models). Any selection criteria or limitations of the samples must be fully described and justified;(v) Statistical methods must consider the predefined endpoints:
(A) Estimates of probabilities of incorrect results must be provided for each endpoint,
(B) Where multiple samples from the same patient are used, statistical analysis must not assume statistical independence without adequate justification, and
(C) The report must provide appropriate confidence intervals for each performance metric;
(vi) Sensitivity and specificity must be characterized across the range of available measurements;
(vii) Agreement of the simulated measure(s) with clinically acceptable measure(s) must be assessed across the full range of measurements;
(viii) Comparison of the measurement performance must be provided across the range of intended image acquisition hardware; and
(ix) If the device uses a cutoff threshold or operates across a spectrum of disease, it must be established prior to validation, and it must be justified as to how it was determined and clinically validated;
(4) Adequate validation must be performed and controls implemented to characterize and ensure consistency (
*i.e.,* repeatability and reproducibility) of measurement outputs:(i) Acceptable incoming image quality control measures and the resulting image rejection rate for the clinical data must be specified, and
(ii) Data must be provided within the clinical validation study or using equivalent datasets demonstrating the consistency (
*i.e.,* repeatability and reproducibility) of the output that is representative of the range of data quality likely to be encountered in the intended use population and relevant use conditions in the intended use environment;(A) Testing must be performed using multiple operators meeting planned qualification criteria and using the procedure that will be implemented in the production use of the device, and
(B) The factors (
*e.g.,* medical imaging dataset, operator) must be identified regarding which were held constant and which were varied during the evaluation, and a description must be provided for the computations and statistical analyses used to evaluate the data;(5) Human factors evaluation and validation must be provided to demonstrate adequate performance of the user interface to allow for users to accurately measure intended parameters, particularly where parameter settings that have impact on measurements require significant user intervention; and
(6) Device labeling must be provided that adequately describes the following:
(i) The device's intended use, including the type of imaging data used, what the device measures and outputs to the user, whether the measure is qualitative or quantitative, the clinical indications for which it is to be used, and the specific population for which the device use is intended;
(ii) Appropriate warnings specifying the intended patient population, identifying anatomy and image acquisition factors that may impact measurement results, and providing cautionary guidance for interpretation of the provided measurements;
(iii) Key assumptions made in the calculation and determination of simulated measurements;
(iv) The measurement performance of the device for all presented parameters, with appropriate confidence intervals, and the supporting evidence for this performance. Per-vessel clinical performance, including where applicable localized performance according to vessel and segment, must be included as well as a characterization of the measurement error across the expected range of measurement for key parameters based on the clinical data;
(v) A detailed description of the patients studied in the clinical validation (
*e.g.,* age, gender, race or ethnicity, clinical stability, current treatment regimen) as well as procedural details of the clinical study (*e.g.,* scanner representation, calcium scores, use of beta-blockers or nitrates); and(vi) Where significant human interface is necessary for accurate analysis, adequately detailed description of the analysis procedure using the device and any data features that could affect accuracy of results.
In combination with the general controls of the FD&C Act. the Coronary Physiologic Simulation Software Device is subject to the following special controls:
{0}------------------------------------------------
Image /page/0/Picture/0 description: The image shows the logo for the U.S. Food & Drug Administration (FDA). The logo consists of two parts: the Department of Health & Human Services seal on the left and the FDA acronym followed by the full name of the agency on the right. The FDA part of the logo is in blue, with the acronym in a solid blue square and the agency name in a lighter blue. The text reads "FDA U.S. FOOD & DRUG ADMINISTRATION".
April 1, 2022
KeyaMed NA Inc. % Kelliann Payne Partner Hogan Lovells US LLP 1735 Market Street, Floor 23 Philladelphia, Pennsylvania 19103
Re: K213657
Trade/Device Name: DEEPVESSEL FFR Regulation Number: 21 CFR 870.1415 Regulation Name: Coronary Vascular Physiologic Simulation Software Device Regulatory Class: Class II Product Code: PJA Dated: March 1, 2022 Received: March 1, 2022
Dear Kelliann Payne:
We have reviewed your Section 510(k) premarket notification of intent to market the device referenced above and have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to legally marketed predicate devices marketed in interstate commerce prior to May 28, 1976, the enactment date of the Medical Device Amendments, or to devices that have been reclassified in accordance with the provisions of the Federal Food, Drug, and Cosmetic Act (Act) that do not require approval of a premarket approval application (PMA). You may, therefore, market the device, subject to the general controls provisions of the Act. Although this letter refers to your product as a device, please be aware that some cleared products may instead be combination products. The 510(k) Premarket Notification Database located at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm identifies combination product submissions. The general controls provisions of the Act include requirements for annual registration, listing of devices, good manufacturing practice, labeling, and prohibitions against misbranding and adulteration. Please note: CDRH does not evaluate information related to contract liability warranties. We remind you, however, that device labeling must be truthful and not misleading.
If your device is classified (see above) into either class II (Special Controls) or class III (PMA), it may be subject to additional controls. Existing major regulations affecting your device can be found in the Code of Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.
Please be advised that FDA's issuance of a substantial equivalence determination does not mean that FDA has made a determination that your device complies with other requirements of the Act or any Federal statutes and regulations administered by other Federal agencies. You must comply with all the Act's
{1}------------------------------------------------
requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part 801); medical device reporting of medical device-related adverse events) (21 CFR 803) for devices or postmarketing safety reporting (21 CFR 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reportingcombination-products); good manufacturing practice requirements as set forth in the quality systems (OS) regulation (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR 4, Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-
542 of the Act); 21 CFR 1000-1050.
Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR Part 807.97). For questions regarding the reporting of adverse events under the MDR regulation (21 CFR Part 803), please go to https://www.fda.gov/medical-device-safety/medical-device-reportingmdr-how-report-medical-device-problems.
For comprehensive regulatory information about mediation-emitting products, including information about labeling regulations, please see Device Advice (https://www.fda.gov/medicaldevices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn (https://www.fda.gov/training-and-continuing-education/cdrh-learn). Additionally, you may contact the Division of Industry and Consumer Education (DICE) to ask a question about a specific regulatory topic. See the DICE website (https://www.fda.gov/medical-device-advice-comprehensive-regulatoryassistance/contact-us-division-industry-and-consumer-education-dice) for more information or contact DICE by email (DICE@fda.hhs.gov) or phone (1-800-638-2041 or 301-796-7100).
Sincerely,
LCDR Stephen Browning Assistant Director Division of Cardiac Electrophysiology, Diagnostics and Monitoring Devices Office of Cardiovascular Devices Office of Product Evaluation and Quality Center for Devices and Radiological Health
Enclosure
{2}------------------------------------------------
510(k) Number (if known)
K213657
#### Device Name
#### DEEPVESSEL FFR
Indications for Use (Describe)
DEEPVESSEL FFR is a coronary physiological simulation software for the clinical quantitative analysis of previously acquired Computed Tomography DICOM data for clinically stable symptomatic patients with coronary artery disease. It provides DVFFR (a CT-derived FFR measurement) computed from static coronary CTA images using deep learning neural networks that encode imaging, structural, and functional characteristics of coronary arteries through learning.
DEEPVESSEL FFR analysis is intended to support the functional evaluation of coronary artery disease.
The results of the analysis are provided to support qualified clinicians to aid in the evaluation and assessment of coronary arteries. DEEPVESSEL FFR results are intended to be used by qualified clinicians in conunction with the patient's clinical history, symptoms, and other diagnostic tests, as well as the clinician's professional judgment.
Type of Use (Select one or both, as applicable)
7 Prescription Use (Part 21 CFR 801 Subpart D)
[ Over-The-Counter Use (21 CFR 801 Subpart C)
# CONTINUE ON A SEPARATE PAGE IF NEEDED.
This section applies only to requirements of the Paperwork Reduction Act of 1995.
### *DO NOT SEND YOUR COMPLETED FORM TO THE PRA STAFF EMAIL ADDRESS BELOW.*
The burden time for this collection of information is estimated to average 79 hours per response, including the time to review instructions, search existing data sources, gather and maintain the data needed and complete and review the collection of information. Send comments regarding this burden estimate or any other aspect of this information collection, including suggestions for reducing this burden, to:
> Department of Health and Human Services Food and Drug Administration Office of Chief Information Officer Paperwork Reduction Act (PRA) Staff PRAStaff(@fda.hhs.gov
"An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB number."
{3}------------------------------------------------
# 510(k) SUMMARY Keya Medical's DEEPVESSEL FFR K213657
## Submitter
KeyaMed NA Inc. 107 Spring Street Seattle, WA 98104, USA
Phone: 1 (206) 508-1036 Contact Person: Xiaoxiao Liu Date Prepared: March 25, 2022
Name of Device: DEEPVESSEL FFR
Classification Name: Coronary Vascular Physiologic Simulation Software
Regulatory Class: Class II
Product Code: PJA
Predicate Device: HEARTFLOW, INC.'s FFRcT V2.0 (K161772)
## Device Description
DEEPVESSEL FFR is a coronary physiological simulation software for the clinical quantitative and qualitative analysis of previously acquired Computed Tomography DICOM data for clinically stable symptomatic patients with coronary artery disease. It estimates FFR values from static coronary CTA images with extracted coronary tree structures using deep learning neural networks. DEEPVESSEL FFR analysis is intended to support the functional evaluation of CAD.
The software processes these images semi-automatically, and it generates a 3D model of the coronay artery tree and computes DVFFR (CT-derived FFR) values. Qualified image analysts interact with the software by providing manual edits to the 3D coronary artery tree segmentations when needed, and oversees outputs along the processing steps. DVFFR analysis results are sent electronically to the physicians via a third-party service portal application.
DVFFR software is independent of imaging equipment, imaging protocols and equipment vendors; the clinical validation study report includes the specific imaging scanner types and imaging acquisition parameters used in the clinical validation of the product.
### Intended Use / Indications for Use
DEEPVESSEL FFR is a coronary physiological simulation software for the clinical quantitative and qualitative analysis of previously acquired Computed Tomography DICOM data for clinically stable
{4}------------------------------------------------
symptomatic patients with coronary artery disease. It provides DVFFR (a CT-derived FFR measurement) computed from static coronary CTA images using deep learning neural networks that encode imaging, structural, and functional characteristics of coronary arteries through learning.
DEEPVESSEL FFR analysis is intended to support the functional evaluation of coronary artery disease. The results of the analysis are provided to support qualified clinicians to aid in the evaluation and assessment of coronary arteries. DEEPVESSEL FFR results are intended to be used by qualified clinicians in conjunction with the patient's clinical history, symptoms, and other diagnostic tests, as well as the clinician's professional judgment.
## Summary of Technological Characteristics
DEEPVESSEL FFR is a software medical device that is designed to be used by qualified image analysts (trained and certified professionals) to analyze coronary CTA images of clinically stable symptomatic patients with CAD. It calculates CT-derived FFR values from static coronary CTA images using deep learning neural networks. DEEPVESSEL FFR analysis is intended to support the functional evaluation for clinical stable CAD patients.
The software generates the DVFFR analysis results in two main steps. The first step generates a 3D coronary artery tree model from the CTA image automatically using deep learning-based segmentation algorithms. Manual corrections of the segmentation results are allowed when necessary to confirm the accuracy of the 3D coronary artery tree segmentation. In the second step, the deep learning framework consists of a multi-layer perceptron network (MLP) and a bidirectional multi-layer recursive neural network (BRNN), which utilize the segmentation results and the CTA image, to estimate semi-continuous FFR values along the coronary artery centerlines. The output of the analysis is a PDF report with detailed DVFFR assessment and branch visualizations, along with a 3D DVFFR tree model where the DVFFR values are mapped on top of the surface model.
DEEPVESSEL FFR and the predicate device have similar technological characteristics, utilizing computational models to generate CT-derived FFR value for interpretation.
{5}------------------------------------------------
| | Subject Device | Predicate Device |
|---------------------|-----------------------------------------|--------------------------------------|
| | DEEPVESSEL FFR 1.0 | HeartFlow FFRCT V2.0 |
| | | (K161772) |
| Manufacturer | Keya Medical | HeartFlow, Inc. |
| Intended | DEEPVESSEL FFR is a coronary | HeartFlow FFRCT is a coronary |
| Use/Indications for | physiological simulation software for | physiologic simulation software for |
| use | the clinical quantitative and | the clinical quantitative and |
| | qualitative analysis of previously | qualitative analysis of previously |
| | acquired Computed Tomography | acquired Computed Tomography |
| | DICOM data for clinically stable | DICOM data for clinically stable |
| | symptomatic patients with coronary | symptomatic patients with coronary |
| | artery disease. It provides DVFFR (a | artery disease. It provides FFRCT, |
| | CT-derived FFR measurement) | a mathematically derived quantity, |
| | computed from static coronary CTA | computed from simulated |
| | images using deep learning neural | pressure, velocity and blood flow |
| | networks that encode imaging, | information obtained from a 3D |
| | structural, and functional | computer model generated from |
| | characteristics of coronary arteries | static coronary CT images. FFRCT |
| | through learning. | analysis is intended to support the |
| | | functional evaluation of coronary |
| | DEEPVESSEL FFR analysis is | artery disease. |
| | intended to support the functional | |
| | evaluation of coronary artery | The results of this analysis are |
| | disease. | provided to support qualified |
| | The results of the analysis are | clinicians to aid in the evaluation |
| | provided to support qualified | and assessment of coronary |
| | clinicians to aid in the evaluation and | arteries. The results of HeartFlow |
| | assessment of coronary arteries. | FFRCT are intended to be used by |
| | DEEPVESSEL FFR results are | qualified clinicians in conjunction |
| | intended to be used by qualified | with the patient's clinical history, |
| | clinicians in conjunction with the with | symptoms, and other diagnostic |
| | the patient's clinical history, | tests, as well as the clinician's |
| | symptoms, and other diagnostic | professional judgment. |
| | tests, as well as the clinician's | |
| | professional judgment. | |
| Intended End User | Clinicians | Clinicians |
| Clinical Condition | Coronary Artery Disease | Coronary Artery Disease |
| Input | Coronary CTA DICOM image data | Coronary CTA DICOM image data |
| Output | 3D Model & Analysis Report | 3D Model & Analysis Report |
# Table 1. Key Feature Comparison
{6}------------------------------------------------
# Performance Data
The following testing have been conducted to demonstrate the substantial equivalence:
Software: Software verification and validation activities were performed according to written procedures and FDA Guidance document Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices. Verification and validation testing confirmed that the predefined acceptance criteria have been fulfilled.
Human Factors Evaluations: Two human factors studies were conducted in accordance with FDA's Guidance Applying Human Factors and Usability Engineering to Medical Devices, 2016. The studies evaluated the critical tasks associated with use of the device for both physicians and analysts. The findings from the study demonstrated that all critical tasks were completed without use error or difficulty. These study results concluded that the DEEPVESSEL reports are safe and effective for the physician user group for the intended uses and use environment. Additionally, the analyst user group were able to safety and accurately use the DVFFR software and generate reports.
Reproducibility/Repeatability Evaluations: Reproducibility (R&R) testing was performed on a group of CT scans with diverse disease conditions and image qualities to evaluate the variation of repeated analyses of DEEPVESSEL FFR with different image analysts (reproducibility) at different days with a washout-period in between to avoid memory effects (repeatability). Testing results met the pre-specified variability metric threshold and thus demonstrated acceptable performance.
Clinical Studies: The software was also validated via a multi-national (US and EU), multicenter clinical validation study with intended patient population to ensure the clinical effectiveness. The primary endpoints of the study were per-vessel sensitivity and specificity of DVFFR to detect ischemic condition comparing with invasive FFR measurement. DVFFR analysis was conducted on a total of 244 patients with 311 target vessels from 8 clinical sites (4 from EU and 4 from US).
At the vessel level, the observed sensitivity of DVFFR was 86.9% with a two-sided lower bound 95% CI of 80.6%, and the observed specificity of DVFFR was 86.7% with a two-sided lower bound 95% Cl of 82.0%. Both 95% Cl lower bounds for sensitivity and specificity exceeded the performance target of 75% and 70%, respectively, as shown in Table 1.
| | Estimate, %<br>(two-sided 95% CI) | Lower Bound of the<br>two-sided 95% CI | Target Rate | Met/Not Met |
|---------------------------------------------------|-----------------------------------|----------------------------------------|-------------|-------------|
| Sensitivity | 86.9%<br>(80.6%-92.7%) | 80.6% | 75% | Met |
| Specificity | 86.7%<br>(82.0%-91.1%) | 82.0% | 70% | Met |
| Positive: measured or estimated FFR values ≤ 0.80 | | | | |
Table 1. Per-vessel sensitivity and specificity of DVFFR
The observed diagnostic accuracy, PPV (positive predictive value) and NPV (negative predictive value) of DVFFR were 86.8% (95% Cl: 83.0%-90.4%), 79.4% (95% Cl: 71.8%-86.2%) and 91.9% (95% CI: 87.7%–95.6%), respectively. The results are summarized in Table 2.
{7}------------------------------------------------
| | Accuracy<br>(two-sided 95% CI) | PPV<br>(two-sided 95% CI) | NPV<br>(two-sided 95% CI) |
|-------|--------------------------------|---------------------------|---------------------------|
| DVFFR | 86.8%<br>(83.0%-90.4%) | 79.4%<br>(71.8%-86.2%) | 91.9%<br>(87.7%-95.6%) |
Table 2. Per-vessel diagnostic accuracy, PPV and NPV of DVFFR
At the patient level, the observed sensitivity, specificity, accuracy, PPV, and NPV were 87.4% (95% Cl: 79.4%-93.1%), 83.7% (95% Cl: 76.5%-89.4%), 85.2% (95% Cl: 80.2%-89.4%), 79.6% (95% Cl: 71.0%-86.6%), and 90.1% (95% Cl: 83.6%-94.6%), respectively, as shown in Table 3.
| | Sensitivity<br>(95% CI) | Specificity<br>(95% CI) | Accuracy<br>(95% CI) | PPV<br>(95% CI) | NPV<br>(95% CI) |
|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------|-------------------------|------------------------|------------------------|------------------------|
| DVFFR | 87.4%<br>(79.4%–93.1%) | 83.7%<br>(76.5%–89.4%) | 85.2%<br>(80.2%–89.4%) | 79.6%<br>(71.0%–86.6%) | 90.1%<br>(83.6%–94.6%) |
| At the patient level, if a patient had more than one ischemic lesion, ICA-FFR value for this patient would be determined as the minimum ICA-FFR measurement from all the coronary arteries. Similarly, patient-level DVFFR value is the minimum DVFFR measurement from all the vessels measured for the patient. | | | | | |
Table 3. Patient-level diagnostic performance of DVFFR
The study demonstrated that DEEPVESSEL FFR yielded good diagnostic performance and met the pre-specified criteria for study success.
# Conclusions
DEEPVESSEL FFR, a coronary physiological simulation software, is substantially equivalent to the predicate device, HeartFlow FFRct V2.0 (K161772). DEEPVESSEL FFR and HeartFlow FFRct V2.0 (K161772) share the same intended use and very similar indications for use, technological characteristics, and principles of operation. The only differences between the subject and predicate devices are the algorithms used to calculate the CT-derived FFR values, and these differences do not raise new questions of safety or effectiveness.
Predicate graph will load when search results are available.
Embedding visualization will load when search results are available.
PDF viewer will load when search results are available.
Loading panels...
Select an item from Submissions
Click any panel, subpart, regulation, product code, or device to see details here.
Section Matches
Results will appear here.
Product Code Matches
Results will appear here.
Special Control Matches
Results will appear here.
Loading collections...
Loading
My Alerts
You will receive email notifications based on the filters and frequency you set for each alert.
Sort by:
Create Alert
Search Filters
Agent Token
Create a read-only bearer token for Claude, ChatGPT, or other agents that can call HTTP APIs.
Copy this now. It will not be shown again.
Connected apps
Apps you authorized through browser sign-in. Disconnecting revokes their access immediately.
Learn the FDA Browser
Two short videos show you everything — or skip straight to the written tutorial if you'd rather read. You can reopen this any time from the Tutorial button in the top bar.
Part 1 — Search, results, and everyday workflows 16 min
Part 2 — Embeddings: the galaxy map 3 min
1. Search: exact and fuzzy
Type a phrase like "coronary artery calcification" into the search box. You get two kinds of results. Exact results match the literal phrase — prefix searches work ("coronary artery calcificati") but suffix searches do not. Fuzzy results match on the meaning and intent of your phrase rather than the exact words, and are sorted by relevance score. Hover over the Exact or Fuzzy badge on any row to see exactly why it matched.
Use the checkboxes above the results to narrow: SaMD keeps only software-only devices, AI / ML keeps only devices with AI.
Exact vs. fuzzy search: what's the difference?
Exact matches on the literal phrase (prefix search works, suffix does not). Fuzzy matches on the meaning and intent of the phrase rather than the exact words. Hover over the badge on any row to see why it matched.
You search "coronary artery calcification" and want only software devices with AI. What two filters do you apply?
Narrow by SaMD (software-only devices), then narrow by AI/ML (devices with AI).
2. The results table
Scroll right in the results table. The intended use is extracted for you — no need to open the PDF. The device story gives a high-level snapshot of what the device does and how it's used. The AI Performance sub-table shows each output name, acceptance criteria, observed values, and development/test dataset descriptions — the same format Innolitics uses for regulatory strategy outputs, and the fastest high-level fingerprint of an AI device. It is AI-generated but has been very reliable in practice.
Where do you find a device's intended use without opening the PDF?
Scroll right in the search results table. The intended use column is extracted for you; no need to dig into the 510(k) summary PDF.
What does the AI Performance sub-table show, and why is it useful?
Output name, acceptance criteria, observed values, development dataset description, and test dataset description. It's the same format we use for regulatory strategy output and Fast 510(k) input, and the fastest high-level fingerprint of an AI device. AI-generated but reliable in practice.
3. Judging fuzzy relevance
Fuzzy results trail off in relevance as you scroll. Use three signals to decide how far down to go: the fuzzy badge explanations, the intended use column, and whether your target output (e.g., Cobb angle) still appears in the AI Performance sub-table. Once it stops appearing, you're past the relevant zone. A top hit with a low score (~0.4) and a stretched explanation is a hint the closest predicates are far away — the project may be headed for De Novo. Note the fuzzy search is a pattern match: it doesn't handle negation ("not") well, and hardware devices can appear — filter by SaMD/AI ML to cut them.
How do you judge how far down fuzzy search results to go?
Use the relevancy signals: the fuzzy badge explanations, the intended use column, and whether the target output (e.g., Cobb angle) still appears in the AI Performance sub-table. Once it stops appearing, results are trailing off in relevancy.
4. Device detail page: chat and citations
Click a device name to open its detail page: device facts on the left, a chat window on the right. Ask something like "Describe the training data". The answer carries little citation bubbles — click one to jump to the highlighted passage in the source PDF, so you can verify every AI answer against the document. There's also a Download PDF button for sharing.
How do you verify an AI chat answer on the device detail page?
Click the citation bubbles to jump to the relevant highlight in the source document.
Reading rule for every project: how many summaries do you read in full?
At least the three most relevant 510(k) or De Novo summaries, in full. After that, use targeted chat questions to confirm your memory quickly. The tool supports this professional habit — it doesn't replace it.
5. Side-by-side comparison
Select multiple rows in the results table (aim for under ~10), then open the PDF Viewer tab. Ask one question — it goes to all selected devices in parallel, each with citations. This is the fastest way to compare and contrast devices: training data, PCCP scope, how they handled adding new scanners, and so on.
What does the side-by-side PDF viewer mode do?
Select multiple devices, open the PDF viewer tab, and ask one question (e.g., "Describe the training data"). It queries all selected devices simultaneously with citations, so you can compare and contrast quickly.
6. Collections
With rows selected, go to the Collections tab and create a labeled collection (e.g., "Cobb Angle Project"). Reload that selection any time — before a client call, pull up the collection and ask questions across all of its devices at once.
How do you save a set of selected devices for later use?
Select the rows, go to the Collections tab, and create a labeled collection (e.g., "Cobb Angle Project"). You can reload the selection anytime and carry it into the PDF viewer and other tabs that support selections.
7. Product codes and the regulations tree
Click a product code in the results to jump to it in the regulations tree — identification text, sibling product codes, and devices you can open in a PDF viewer on the right. Click a regulation number to see its identification, special controls, and related product codes. You can also search by product code or regulation number at the top of the tree. Always read the special controls if any exist for your device — it broadens your search and sharpens pre-kickoff research.
What can you do from the regulations tree view?
Browse product codes and regulation numbers, read the identification text and special controls, browse sibling product codes, open device PDFs on the right, and search by product code or regulation number at the top of the tree.
8. Chart view
Click Show Chart and segment by regulation number (or product code) to see which regulations dominate your result set. Clicking a regulation takes you into the regulations tree. Great for spotting that most matches are, say, hardware laparoscopic devices — a cue to go back and filter.
How do you see which regulations dominate a search result set?
Click "Show Chart" and segment by Regulation Number. Clicking a regulation takes you to the regulations tree.
9. The predicate graph
Open the Predicates tab for a family-tree view of predicate relationships. Click a node to trace its parents and children; selections from search carry over pre-selected. Commonly predicated devices are worth reading — a lot of people predicated them for a reason. The visual lineage is also handy on client calls, e.g. to show how a predicate family evolved and justify why your predicate still holds.
In the predicate graph, why are commonly predicated devices worth reading?
A lot of people predicated them for a reason. Clicking a node traces parents and children, and selections from search carry over pre-selected.
10. Embeddings: the galaxy map
The Embeddings tab plots every matching document in a 2-D "galaxy map" where semantically similar devices cluster together. Hover or click clusters to explore, and let AI label the clusters for you. Embeddings beat product codes for grouping: two devices can carry different product codes (LLZ vs. QIH) yet do the same thing — the embedding captures the meaning of the intended use and device story. This is also exactly how retrieval-augmented generation (RAG) works under the hood, and it makes a great visual on client calls.
Try it yourself
Head to the search page and work through a few of these AI/ML fuzzy searches to build intuition: perivascular fat on CT · aortic valve calcification opportunistic screening on noncontrast CT · breast cancer prediction on digital pathology slides · autism detection · gestational age prediction · a hearing aid that can also detect a pulse · foundation model based analysis of ECG · large language models · penetration test. Watch how the relevance scores, intended use, and AI Performance tables tell you when results stop being meaningful.
