VERIGENE SYSTEM, F2 NUCLEIC ACID TEST, F5 NUCLEIC ACID TEST, AND MTHFR NUCLEIC ACID TEST
K070597 · Nanosphere, Inc. · NSU · Oct 11, 2007 · Clinical Chemistry
Device Facts
Record ID
K070597
Device Name
VERIGENE SYSTEM, F2 NUCLEIC ACID TEST, F5 NUCLEIC ACID TEST, AND MTHFR NUCLEIC ACID TEST
Applicant
Nanosphere, Inc.
Product Code
NSU · Clinical Chemistry
Decision Date
Oct 11, 2007
Decision
SESE
Submission Type
Traditional
Regulation
21 CFR 862.2570
Device Class
Class 2
Indications for Use
The Verigene® F5 Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (G to A at position 1691; also known as Factor V Leiden) of the human Factor V gene (F5; Coagulation Factor V gene) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System. The Verigene® F2 Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (G to A at position 20210 of the human Factor II gene (F2; prothrombin gene)) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System. The Verigene® MTHFR Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (C to T at position 677) of the human 5, 10-methylenetetrahydrofolate reductase gene (MTHFR) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System.
Device Story
Verigene System is a bench-top molecular diagnostics workstation for automated genotyping of DNA samples. Input: isolated genomic DNA loaded into disposable Test Cartridges. Operation: hybridization of target DNA to synthetic gene-specific oligonucleotide capture strands on cartridge substrate; formation of 'sandwich' complex with mediator oligonucleotide; binding of gold nanoparticle probes; signal enhancement via two-part reagent. Reader illuminates nanoparticles; photosensor measures relative brightness of wild-type vs. mutant spots. Output: genotype result based on relative brightness. Used in clinical laboratories by trained professionals. Benefits: automates complex molecular analysis; provides rapid (approx. 1.5 hours) genotyping to aid diagnosis of hereditary thrombophilia factors.
Clinical Evidence
Method comparison study using 287 clinical samples (68% from 'rule-out thrombophilia' patients) compared Verigene results to bi-directional DNA sequencing. Results showed 100% Positive Percent Agreement (95% CI: 98.9-100.0%) and 100% Negative Percent Agreement (95% CI: 98.9-100.0%). No clinical data provided; bench testing only.
Technological Characteristics
System uses disposable cartridges with synthetic oligonucleotide capture strands and gold nanoparticle probes. Detection via evanescent wave light scatter. Requires isolated genomic DNA (40-400 ng/μL). Connectivity: Up to 8 processors per reader. Calibration required for temperature control system. Software performs automated on-line quality monitoring of fluidics, instrument, and reagent integrity.
Indications for Use
Indicated for patients with suspected thrombophilia requiring detection and genotyping of Factor V Leiden (G1691A), Factor II (G20210A), or MTHFR (C677T) mutations using isolated genomic DNA from whole blood.
Regulatory Classification
Identification
Instrumentation for clinical multiplex test systems is a device intended to measure and sort multiple signals generated by an assay from a clinical sample. This instrumentation is used with a specific assay to measure multiple similar analytes that establish a single indicator to aid in diagnosis. Such instrumentation may be compatible with more than one specific assay. The device includes a signal reader unit, and may also integrate reagent handling, hybridization, washing, dedicated instrument control, and other hardware components, as well as raw data storage mechanisms, data acquisition software, and software to process detected signals.
Special Controls
*Classification.* Class II (special controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 862.9. The special control is FDA's guidance document entitled “Class II Special Controls Guidance Document: Instrumentation for Clinical Multiplex Test Systems.” See § 862.1(d) for the availability of this guidance document.
The special control is FDA's guidance document entitled "Class II Special Controls Guidance Document: Instrumentation for Clinical Multiplex Test Systems."
Roche Diagnostics Corporation Factor V Leiden Kit (K033607)
Roche Diagnostics Corporation Factor II (Prothrombin) G20210A Kit (K033612)
Submission Summary (Full Text)
{0}
1
# 510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION DECISION SUMMARY
A. 510(k) Number:
k070597
B. Purpose for Submission:
New device
C. Measurand:
Factor II (Prothrombin)
Factor V Leiden
5, 10-Methylenetetrahydrofolate reductase (MTHFR)
D. Type of Test:
Genotype
E. Applicant:
Nanosphere INC
F. Proprietary and Established Names:
Verigene® F5 Nucleic Acid Test
Verigene® F2 Nucleic Acid Test
Verigene® MTHFR Nucleic Acid Test
G. Regulatory Information:
1. Regulation section:
21 CFR 864.7280 Factor V Leiden DNA mutation detection systems
2. Classification:
Class II
3. Product code:
NPR, Test, Factor II G20210A mutations, genomic DNA PCR
NPQ, Test, Factor V Leiden mutations, genomic DNA PCR
OMM, Test, 5, 10-Methylenetetrahydrofolate reductase mutations, genomic DNA PCR
NSU, Instrumentation for Clinical Multiplex Test Systems
4. Panel:
Hematology (81); Chemistry (75)
H. Intended Use:
1. Intended use(s):
The Verigene® F5 Nucleic Acid Test is an *in vitro* diagnostic for the detection and genotyping of a single point mutation (G to A at position 1691; also known as Factor V Leiden) of the human Factor V gene (F5; Coagulation Factor V gene) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System.
The Verigene® F2 Nucleic Acid Test is an *in vitro* diagnostic for the detection and genotyping of a single point mutation (G to A at position 20210 of the human Factor II gene (F2; prothrombin gene)) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System.
{1}
The Verigene® MTHFR Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (C to T at position 677) of the human 5, 10-methylenetetrahydrofolate reductase gene (MTHFR) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System.
2. Indication(s) for use:
Same as Intended Use
3. Special conditions for use statement(s):
For Prescription Use Only.
4. Special instrument requirements:
Verigene® System
I. Device Description:
The Verigene® System consists of two instruments, the Verigene® Processor and the Verigene® Reader. It utilizes single-use, disposable test Cartridges to process and genotype multiple genes in a DNA sample in ~1 ½ hrs. Up to 8 Verigene® Processors may be connected to a single Verigene® Reader. Each processor contains 4 hybridization modules, and the modules can simultaneously run different tests.
J. Substantial Equivalence Information:
1. Predicate device name(s):
Roche Diagnostic Corp. Factor V Leiden Kit
Roche Diagnostic Corp. Factor II (Prothrombin) G20210A Kit.
2. Predicate 510(k) number(s):
k033607, k033612
3. Comparison with predicate:
| Similarities | | |
| --- | --- | --- |
| Item | Device | Predicate |
| Intended use | The Verigene® F5 Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (G to A at position 1691; also known as Factor V Leiden) of the human Factor V gene (F5; Coagulation Factor V gene) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System. | The Factor V Leiden Kit is an in vitro diagnostic test for the detection and genotyping of a single point mutation (G to A at position 1691) of the human Factor V gene, from DNA isolated from human whole peripheral blood. The Factor V Leiden Kit is indicated as an aid to diagnosis in the evaluation of patients with suspected thrombophilia |
| Intended use | The Verigene® F2 Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (G to A at position | The Factor II(Prothrombin) G20210OA Kit is an in vitro diagnostic test for the detection and genotyping |
{2}
| Similarities | | |
| --- | --- | --- |
| Item | Device | Predicate |
| | 20210of the human Factor II gene (F2; prothrombin gene) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System. | of a single point mutations (G to A at position 202 10) of the human Factor II gene, from DNA isolated from human whole peripheral blood. The Factor 1I (Prothrombin) G20210A Kit is indicated as an aid to diagnosis in the evaluation of patients with suspected thrombophilia |
| Specimen Type | Peripheral whole blood | same |
| Differences | | |
| --- | --- | --- |
| Item | Device | Predicate |
| Intended use | The Verigene® MTHFR Nucleic Acid Test is an in vitro diagnostic for the detection and genotyping of a single point mutation (C to T at position 677) of the human 5, 10-methylenetetrahydrofolate reductase gene (MTHFR) in patients with suspected thrombophilia, from isolated genomic DNA obtained from whole blood samples. The test is intended to be used on the Verigene® System. | None |
| Detection method | Signal amplification Chemical amplification of reporter signal | DNA amplification (PCR) |
| Sample size | 25 μL | 10-20 μL in glass capillaries |
| Detection procedure | Single-image sensor where nanoparticles are illuminated using a fixed-wavelength light source | Optical detection of stimulated fluorescence |
| Detection Chemistry | SNP discrimination via oligonucleotide probes; diction via evanescent wave light scatter with nanoparticles | Paired hybridization probes using fluorescence resonance energy transfer (FRET) |
{3}
4
| Differences | | |
| --- | --- | --- |
| Item | Device | Predicate |
| Analysis time | 1.5 hrs | Detection occurs at defined intervals during PCR cycle and can e viewed in real-time |
K. Standard/Guidance Document Referenced (if applicable):
Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices
Class II Special Controls Guidance Document: Factor V Leiden DNA Mutation Detection Systems
CLSI EP5-A2; Evaluation of Precision Performance of Quantitative Measurement Methods
L. Test Principle:
The analysis sequence is the same for each of the three tests. After extracted and purified DNA, mixed with hybridization buffer, is loaded into the sample well of the test cartridge, it is ready for processing. An internal barcode reader reads the cartridge ID and sends the information to the Verigene® Reader. The reader establishes the hybridization parameters and starts the hybridization process.
The genotyping process occurs with a hybridization of the target analyze to a synthetic gene-specific oligonucleotide capture strand on the test cartridge’s substrate. A synthetic mediator target-specific oligonucleotide is included with the test-specific sample buffer to form a hybridization” sandwich” with the gene sequence of interest. After washing, a probe composed of a gold nanoparticle with covalently bound oligonucleotide complementary to a sequence on the intermediate oligonucleotide, is introduced. A signal enhancement reagent is added to the hybridization chamber and reacts with the gold nanoparticle to amplify the signal for the Verigene® Reader scanning and analysis.
M. Performance Characteristics (if/when applicable):
1. Analytical performance:
a. Precision/Reproducibility:
The reproducibility testing included 4 studies, each using a different DNA samples extracted from whole blood. In the first study, each of the three test sites ran the same sample in duplicate.
| Sites | F5 | F2 | MTHFR |
| --- | --- | --- | --- |
| 1 | HET | HET | HET |
| | HET | HET | HET |
| 2 | HET | HET | HET |
| | HET | HET | HET |
| 3 | HET | HET | HET |
| | HET | HET | HET |
In the second study, one operator at Site 1 analyzed the same sample in duplicate each
{4}
day for three days.
| Days | F2 | F5 | MTHFR |
| --- | --- | --- | --- |
| 1 | WT | HET | WT |
| | WT | HET | WT |
| 2 | WT | HET | WT |
| | WT | HET | WT |
| 3 | WT | HET | WT |
| | WT | HET | WT |
In the third study, three operators at Site 2 each analyzed the same sample in duplicate.
| Operators | F2 | F5 | MTHFR |
| --- | --- | --- | --- |
| 1 | WT | HET | HET |
| | WT | HET | HET |
| 2 | WT | HET | HET |
| | WT | HET | HET |
| 3 | WT | HET | HET |
| | WT | HET | HET |
In the fourth study, one operator at Site 3 analyzed the same sample in duplicate using three lots of reagents.
| Reagent Lots | F2 | F5 | MTHFR |
| --- | --- | --- | --- |
| 1 | HET | WT | HET |
| | HET | WT | HET |
| 2 | HET | WT | HET |
| | HET | WT | HET |
| 3 | HET | WT | HET |
| | HET | WT | HET |
b. Linearity/assay reportable range: NA
c. Traceability, Stability, Expected values (controls, calibrators, or methods): NA
d. Detection limit: In 40 genomic samples diluted to $40\mathrm{ng} / \mu \mathrm{L}$ the total call rate for all three assays was $82\%$ . There were no mis-calls.
e. Analytical specificity: To simulate sample cross-contamination, a sample of known genotype was spiked with various amounts of a sample with a known but different genotype. The first sample consisted of a F2 (wild-type); F5 (mutant); MTHFR (wild type) and the
{5}
contaminating sample's genotype was: F2 (mutant); F5 (wild-type); MTHFR (hetero). The percent contaminating sample ranged from 0-50%. The results indicated that neither the call rate nor accuracy in a homozygous sample was affected in the presence of up to 10% contamination by volume of a sample with the opposite homozygous genotype. At least 20% contamination by volume of a sample with a heterozygous genotype is required to affect the call rate.
A study was performed to assess the interference of heparin, hemoglobin or magnetic beads in a purified DNA sample. Interference was observed in the 5X magnetic bead sample where the F5 signal was slightly lower than the control. The genotyping ration results showed slight variations but none led to an increase in no-calls or a mis-call.
f. Assay cut-off: NA
2. Comparison studies:
a. Method comparison with predicate device:
Accuracy (percent agreement) was determined by comparison to bi-directional DNA sequencing:
A total of 287 samples, sixty-eight percent (68%) from patients undergoing "rule-out thrombophilia" testing, were analyzed at three sites using the Verigene® F5 / F2 / MTHFR Nucleic Acid Tests and by bi-directional sequencing analysis at an independent reference laboratory. All purified DNA samples were from whole blood collected using EDTA as the anticoagulant. Data comparison details for F5, F2, and MTHFR are shown in tables 3-5 below:
Table 3. Verigene® F5 Nucleic Acid Test method comparison results
| | Verigene® F5 Nucleic Acid Test | | | |
| --- | --- | --- | --- | --- |
| Sequence analysis* | | Wild-type (wt) | Heterozygous (het) | Mutant (mut) |
| | wt | 253 | 0 | 0 |
| | het | 0 | 22 | 0 |
| | mut | 0 | 0 | 7 |
Table 4. Verigene® F2 Nucleic Acid Test method comparison results
| | Verigene® F2 Nucleic Acid Test | | | |
| --- | --- | --- | --- | --- |
| Sequence analysis* | | Wild-type (wt) | Heterozygous (het) | Mutant (mut) |
| | wt | 258 | 0 | 0 |
| | het | 0 | 9 | 0 |
| | mut | 0 | 0 | 5 |
Table 5. Verigene® MTHFR Nucleic Acid Test method comparison results
{6}
| | Verigene® MTHFR Nucleic Acid Test | | | |
| --- | --- | --- | --- | --- |
| Sequence analysis* | | Wild-type (wt) | Heterozygous (het) | Mutant (mut) |
| | wt | 125 | 0 | 0 |
| | het | 0 | 117 | 0 |
| | mut | 0 | 0 | 26 |
*Two DNA samples did not return complete bi-directional sequencing results and, therefore, were excluded from the data. Specifically, one sample did not produce results for any genotype reads (i.e., $F2$ , $F5$ , and MTHFR) and one sample returned no results for the $F2$ genotype. These four genotype reads represent $0.7\%$ of the total possible calls in the method comparison study. There were no samples excluded due to discrepancies between sequenced results.
The Verigene® F5 / F2 / MTHFR Nucleic Acid Tests demonstrated $100\%$ Positive Percent Agreement $(95\% \mathrm{CI} = 98.9$ to $100.0\%)$ and $100\%$ Negative Percent Agreement $(95\% \mathrm{CI} = 98.9$ to $100.0\%)$ in these studies based on data from calls made.
b. Matrix comparison: NA
3. Clinical studies:
a. Clinical Sensitivity: NA
b. Clinical specificity: NA
c. Other clinical supportive data (when a. and b. are not applicable): NA
4. Clinical cut-off: NA
5. Expected values/Reference range: NA
N. Instrument Name:
Verigene® System
O. System Descriptions:
1. Modes of Operation:
Closed system
2. Software:
FDA has reviewed applicant's Hazard Analysis and software development processes for this line of product types:
Yes ☐ or No ☑
3. Specimen Identification:
Barcode reader
4. Specimen Sampling and Handling:
DNA should be extracted using a DNA extraction method that provides DNA with the following characteristics:
Average purity: Optimal mean: 1.89 (typical range: 1.60 to 2.00 $(\mathrm{A}_{260} / \mathrm{A}_{260})$
{7}
Concentration: 40 ng/μL to 400 ng/μL
DNA elution: water or Tris-EDTA
5. Calibration:
The temperature control system on the Verigene® Processor device is the only component that requires calibration. This is performed at the time of manufacture or by a Nanosphere service technician.
6. Quality Control:
1. Automated on-line quality misprints that monitor instrument functionality, software performance, fluidics, test conditions, reagent integrity, and procedural steps in each assay each time a test is performed
2. Liquid materials
P. Other Supportive Instrument Performance Characteristics Data Not Covered In the "Performance Characteristics" Section above:
Q. Proposed Labeling:
The labeling is sufficient and it satisfies the requirements of 21 CFR Part 809.10.
R. Conclusion:
The submitted information in this premarket notification is complete and supports a substantial equivalence decision.
8
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.