Retrospective clinical specimens were used in a comparison study to demonstrate that the modified internal and positive controls did not alter the assay's performance compared to the predicate device.
Retrospective clinical specimens; Archived samples; Comparison study
Clinical Evidence
Study Design
Population
Comparator
Key Endpoints
Retrospective comparison study
Archived nasopharyngeal swab (NPS) specimens; Sample Size: 330 retrospective samples (out of 432 total samples including contrived)
Current Prodesse ProFlu+ Assay (RIC)
Percent Positive Agreement and Percent Negative Agreement
Indications for Use
The Prodesse ProFlu+ Assay is a multiplex real-time PCR in vitro diagnostic test for the qualitative detection and differentiation of Influenza A virus, Influenza B virus, and Respiratory Syncytial Virus (RSV) nucleic acids in nasopharyngeal (NP) swab specimens from individuals with signs and symptoms of respiratory tract infection. The ProFlu+ Assay is intended to aid in the differential diagnosis of Influenza A, Influenza B, and RSV viral infections in humans. Negative results do not preclude influenza or RSV infection and should not be used as the sole basis for diagnosis, treatment or other patient management decisions. The ProFlu+ Assay is intended for use by professional laboratory personnel.
Device Story
ProFlu+ Assay is a multiplex RT-PCR in vitro diagnostic test; detects/discriminates Influenza A, Influenza B, and RSV nucleic acids. Input: nasopharyngeal swab specimens in viral transport medium. Process: nucleic acid isolation/purification via MagNA Pure LC or NucliSENS easyMAG; reverse transcription and amplification using Cepheid SmartCycler II. Taqman chemistry utilizes 5'-3' exonuclease activity of Taq polymerase to cleave dual-labeled probes, generating fluorescent signals (FAM, TET, Texas Red, Cy5). Universal Internal Control (UIC) monitors for inhibitors. Output: qualitative detection of viral targets monitored by fluorescent intensity during PCR cycles. Used in clinical laboratories by trained personnel. Results aid differential diagnosis; negative results do not exclude infection; positive results do not rule out co-infections. Benefits: rapid identification of respiratory viral pathogens to guide clinical management.
Clinical Evidence
Retrospective clinical comparison study demonstrated the modified assay with Universal Internal Control meets performance claims of the original device. Analytical studies confirmed limit of detection for H3N2v and H7N9 strains. Stability studies validated performance after 10 freeze-thaw cycles for enzymes.
Technological Characteristics
Multiplex real-time RT-PCR. Analyte: RNA. Instrumentation: bioMérieux NucliSENS easyMAG or Roche MagNA Pure and Cepheid SmartCycler II. Controls: Universal Internal Control (RNA IVT + DNA plasmid) and pooled positive controls. Stability: M-MLV Reverse Transcriptase and RNase Inhibitor II (10 freeze-thaw cycles).
Indications for Use
Indicated for symptomatic patients requiring differential diagnosis of Influenza A, Influenza B, and RSV viral infections via nasopharyngeal swab specimens. Not intended for Influenza C detection.
Regulatory Classification
Identification
A respiratory viral panel multiplex nucleic acid assay is a qualitative in vitro diagnostic device intended to simultaneously detect and identify multiple viral nucleic acids extracted from human respiratory specimens or viral culture. The detection and identification of a specific viral nucleic acid from individuals exhibiting signs and symptoms of respiratory infection aids in the diagnosis of respiratory viral infection when used in conjunction with other clinical and laboratory findings. The device is intended for detection and identification of a combination of the following viruses:(1) Influenza A and Influenza B; (2) Influenza A subtype H1 and Influenza A subtype H3; (3) Respiratory Syncytial Virus subtype A and Respiratory Syncytial Virus subtype B; (4) Parainfluenza 1, Parainfluenza 2, and Parainfluenza 3 virus; (5) Human Metapneumovirus; (6) Rhinovirus; and (7) Adenovirus.
Special Controls
*Classification.* Class II (special controls). The special controls are:(1) FDA's guidance document entitled “Class II Special Controls Guidance Document: Respiratory Viral Panel Multiplex Nucleic Acid Assay;”
(2) For a device that detects and identifies Human Metapneumovirus, FDA's guidance document entitled “Class II Special Controls Guidance Document: Testing for Human Metapneumovirus (hMPV) Using Nucleic Acid Assays;” and
(3) For a device that detects and differentiates Influenza A subtype H1 and subtype H3, FDA's guidance document entitled “Class II Special Controls Guidance Document: Testing for Detection and Differentiation of Influenza A Virus Subtypes Using Multiplex Nucleic Acid Assays.” See § 866.1(e) for the availability of these guidance documents.
{0}
SPECIAL 510(k): Device Modification Review Memorandum
To: Hologic, Inc. (Gen-Probe Prodesse, Inc.)
RE: K132129
This 510(k) submission contains information/data on modifications made to the SUBMITTER'S own Class II devices requiring 510(k). The following items are present and acceptable:
1. The name and 510(k) number of the SUBMITTER'S previously cleared device:
Prodesse ProFlu™+ Assay
510(k) number: K110968
2. Submitter's statement that the INDICATION/INTENDED USE of the modified device as described in its labeling HAS NOT CHANGED along with the proposed labeling which includes instructions for use and package labeling.
3. A description of the device MODIFICATION(S) to demonstrate that the FUNDAMENTAL SCIENTIFIC TECHNOLOGY of the modified device has not changed.
The 510k submission contained modifications to the Internal Control and Positive Controls as well as expanded Reactivity table to include two additional strains of Influenza A virus, Influenza A/Indiana/10/2011 and Influenza A/Anhui/1/2013. The modifications are summarized as follows:
a. Outsourcing of the manufacturing of the Internal Control and subsequent minor changes to vector sequence;
The current Internal Control (RIC) in ProFlu+ Assay contains a RNA in vitro transcript (IVT). The new Universal Internal Control (UIC-A) will contain a RNA in vitro transcript (IVT) and a DNA plasmid to allow users to perform one nucleic acid extraction and test with any combination of the Pro+ Series Assays including ProFlu+, ProhMPV+, ProParaflu+, ProFAST+, and ProAdeno+. Due to the different vector being used in the Universal Internal Control (UIC-A), a minor change was made to the 5' and 3' ends of the UIC-A sequence.
The concentration of the RNA IVT in the Universal Internal Control (UIC-A) is the same as in the current Internal RNA Control (RIC). Handling of Universal Internal Control is identical to that of the current Internal RNA Control (RIC) included in the ProFlu+ Assay.
b. Outsourcing of the manufacturing of the Positive Control leading to minor changes in the vector sequence, changes to control format and concentration;
The current Positive Controls consist of 4 individual controls for Influenza A (HCT75), Influenza B (HCT76), RSV A (HCT77) and RSV B (JCT78), respectively. The new Positive Controls contain a pooled positive control for Influenza A, Influenza B and RSV A, and a RSV B Control.
- Due to the change in vector, a minor change was made to the 5' and 3' ends of the Control sequences.
- The handling of the Positive Controls for the ProFlu+ Assay will be changed to eliminate the customer dilution that occurs immediately prior to RT-PCR setup, effectively raising the testing concentration one log.
{1}
Page 2 of 6
c. Revised reactivity table to include two additional strains of Influenza A, Influenza A/Indiana/10/2011 (H3N2v) and Influenza A/Anhui/1/2013 (H7N9); The Influenza A/H3N2v can be detected at $10^{2}\mathrm{TCID}_{50} / \mathrm{mL}$ and Influenza A/H7N9 RNA can be detected at $0.02\mathrm{pg} / \mu \mathrm{L}$ .
d. Change in Stability Claims.
The stability study demonstrated that the intermediate stock of the Universal Internal Control can stand up to 2 freeze-thaw cycles, the performance of ProFlu+ Supermix can stand up to 5 freeze-thaw cycles (same as the current stability claim), and M-MLV Reverse Transcriptase and RNase Inhibitor II can stand up to 10 freeze-thaw cycles (increased from the current 5 freeze-thaw cycles)
4. Comparison Information (similarities and differences) to applicant's legally marketed predicate device including, labeling, intended use, and physical characteristics.
| Similarities | | |
| --- | --- | --- |
| Element | Modified Prodesse ProFlu+ Assay | Current Prodesse ProFlu+ Assay (K110968) |
| Organisms Detected | Same | Influenza A virus, Influenza B virus, Respiratory Syncytial Virus |
| Analyte | Same | RNA |
| Technological Principles | Same | Multiplex nucleic acid amplification |
| Specimen Types | Same | Nasopharyngeal Swab |
| User Complexity | Same | High |
| Sample Preparation Method | Same | Up front sample processing is required to extract nucleic acid. |
| Instrumentation | Same | bioMérieux NucliSENS easyMAG or Roche MagNA Pure and Cepheid SmartCycler II Instrument |
| Time to result | Same | Approximately 4 hours |
| Controls | Same | Internal control in each sample. External control processed with each batch of samples. (see below for differences) |
| Differences | | | |
| --- | --- | --- | --- |
| Element | | Modified ProFlu+ Assay | Current Prodesse ProFlu+ Assay |
| Controls | Internal | Universal Internal Control - Contains DNA plasmid in addition to RNA IVT Control Stocks outsourced - Change in manufacturer leading to change in control vectors and minor sequence change at the 5' and 3' ends of RNA IVT | Internal RNA Control - Contains RNA IVT Control stocks manufactured in house |
{2}
Page 3 of 6
| | Positive | • Pooled Influenza A/ Influenza B/RSV A RNA Control and RSV B RNA Control
• Control stocks outsourced.
- Change in manufacturer leading to change in control vectors and minor sequence changes at the 5' and 3' ends of RNA IVTs
• PC does not require dilution;
- PC is provided as “at use concentration” | • Four individual positive controls (Influenza A RNA Control, Influenza B RNA Control, RSV A RNA Control and RSV B RNA Control)
• Control stocks manufactured in house
• End user must dilute PC 1:10 prior to use for RT-PCR |
| --- | --- | --- | --- |
| Reactivity | Influenza A/ Indiana/10/2011 (H3N2v)
Influenza A/ Anhui/1/2013 (H7N9)* | • 10^{2} TCID_{50}/mL
• 0.02 pg/μL | none |
| Stability (Freeze-thaw Cycle) | M-MLV Reverse Transcriptase
RNase Inhibitor II | • 10 cycles
• 10 cycles | • 5 cycles
• 5 cycles |
*Although this test has been shown to detect A/Anhui/1/2013 H7N9 virus RNA and influenza A/ Indiana/10/2011 H3N2v cultured from positive human respiratory specimens, the performance characteristics of this device with clinical specimens that are positive for H7N9 or H3N2v influenza viruses have not been established. The Prodesse ProFlu+™ Assay can distinguish between influenza A and B viruses, but it cannot differentiate influenza A subtypes.
# 5. A Design Control Activities Summary:
a. To demonstrate that the modifications in Controls do not change the assay performance, Analytical Studies and a Comparison Study were conducted.
- Analytical Performances:
- Analytical Sensitivity Confirmation
LoD, which was established in K110968 in 2011, was confirmed for Influenza A, Influenza B, RSV A and RSV B using one strain of each virus when tested with the UIC-A and modified Positive Controls side by side with the current RIC and Positive Controls. The confirmed LoDs are as follows:
Influenza A 1X 10<sup>2</sup> TCID<sub>50</sub>/mL
Influenza B 1X 10<sup>1</sup> TCID<sub>50</sub>/mL
RSV A 1X 10<sup>1</sup> TCID<sub>50</sub>/mL
RSV B 1X 10<sup>2</sup> TCID<sub>50</sub>/mL
- IC Interference Study
The IC Interference Study demonstrated that the new control, UIC-A, did not inhibit the detection of target organisms at levels close to LoD.
- Sample Stability Study
The study demonstrated that the stability of the samples would not be affected by a change in the internal control.
- Extractor Equivalency Studies
{3}
The equivalency of nucleic acid extraction methods between the bioMérieux NucliSENS easyMAG automated extractor and Roche MagNA Pure LC extractor were evaluated by spiking the cultured and tittered strain of Influenza A into a negative nasopharyngeal swab (NPS) matrix pool at the confirmed LoD concentration. The study demonstrated the equivalency between the two extraction methods.
# Comparison Study:
The comparison study was conducted for all Pro+ Series Assays including ProFlu+, ProhMPV+, ProParaflu+, ProFAST+, and ProAdeno+ testing 366 positive samples and 66 negative samples. Among the 366 positive samples, 330 were retrospective pre-selected archived NPS specimens with 30 positive samples per target (11 targets total) and 36 were contrived samples, generated by spiking individual negative retrospective NPS samples with whole organism (Influenza A/Seasonal H1 or Parainfluenza 2). Each sample was split into 3 aliquots; one aliquot was tested using the current Internal RNA Control (RIC), one aliquot was tested using the Universal Internal Control (UIC-A), and one aliquot was tested using the current Universal Internal Control (UIA-P) for ProAdeno+ Assay. All samples were then split into 72 panels with 6 samples per panel, extracted and tested by four different operators. Half of the panel samples were extracted using the bioMérieux NucliSENS easyMAG method and the other half using the Roche MagNA Pure LC method. Of the 432 samples utilized in the study, 21 samples were removed from analysis due to the invalid controls or incomplete test results. The results for ProFlu+ Assay are summarized in the following tables:
| ProFlu+ Assay Influenza A Results | | | | | |
| --- | --- | --- | --- | --- | --- |
| | | Samples with RIC | | Total | |
| Samples with UIC-A | | Positive | Negative | Total | Comments |
| | Positive | 116 | 1* | 117 | Percent Positive Agreement 99.2% (95.3% - 99.9%) 95% CI |
| | Negative | 1** | 293 | 294 | Percent Negative Agreement 99.7% (98.1% - 99.9%) 95% CI |
| Total | | 117 | 294 | 411 | |
*Contrived sample, negative NPS spiked with Influenza A/Seasonal H1
**Sample Influenza B positive with original source laboratory method (culture)
| ProFlu+ Assay Influenza B Results | | | | | |
| --- | --- | --- | --- | --- | --- |
| | | Samples with RIC | | Total | |
| Samples with UIC-A | | Positive | Negative | Total | Comments |
| | Positive | 26 | 1* | 27 | Percent Positive Agreement 100% (87.1% - 100.0%) 95% CI |
| | Negative | 0 | 384 | 384 | Percent Negative Agreement 99.7% (98.5% - 100.0%) 95% CI |
| Total | | 26 | 385 | 411 | |
*Contrived sample, negative NPS spiked with Parainfluenza 2
{4}
Page 5 of 6
| ProFlu+ Assay RSV Results | | | | | |
| --- | --- | --- | --- | --- | --- |
| | | Samples with RIC | | Total | |
| Samples with UIC-A | | Positive | Negative | Total | Comments |
| | Positive | 71 | 0 | 71 | Percent Positive Agreement 97.3% (90.6% - 99.3%) 95% CI |
| | Negative | 2* | 338 | 340 | Percent Negative Agreement 100% (98.9% - 100.0%) 95% CI |
| Total | | 73 | 338 | 411 | |
*One contrived sample (negative NPS spiked with Influenza A/Seasonal H1) and one sample Parainfluenza 2 positive with original source method (Luminex RVP).
The results of the analytical studies and the clinical study confirmed the original performance claims of the ProFlu+ Assay and demonstrated that assay performance was not affected by the incorporation of the modified Universal Internal Control (UIC-A) and Positive Controls. The ProFlu+ Assay package insert has been updated to reflect the changes in the controls.
b. To assess the reactivity of the ProFlu+ Assay with influenza A(H3N2v) virus and influenza A(H7N9) virus, a cultured and tittered strain of H3N2v and purified genomic RNA isolated from a strain of H7N9 were diluted in series to near the assay cutoff. RNA isolated from Flu A/H7N9 instead of a cultured and tittered Flu A/H7N9 virus was used in the reactivity study due to the requirement of biosafety Level 3 unavailable at Gen-Probe Prodesse. The study results showed that the ProFlu+ Assay can detect Influenza A/H3N2v at $10^{2}\mathrm{TCID}_{50} / \mathrm{mL}$ and Influenza A/H7N9 RNA at $0.02\mathrm{pg} / \mu \mathrm{L}$.
Although this test has been shown to detect A/Anhui/1/2013 H7N9 RNA and influenza A/ Indiana/10/2011 (H3N2v) virus cultured from positive human respiratory specimens, the performance characteristics of this device with clinical specimens that are positive for H7N9 or H3N2v influenza viruses have not been established. The Prodesse ProFlu™+ Assay can distinguish between influenza A and B viruses, but it cannot differentiate influenza A subtypes.
The ProFlu+ Assay package insert has been updated to include the revised reactivity table.
c. To assess the stability of the Universal Internal Control (UIA) and new Positive Controls, an accelerated Stability study for the Controls stored at $-70^{\circ}\mathrm{C}$ and a freeze-thaw stability study were conducted with one lot of each assay component. The studies demonstrated that the UIC-A and modified Positive Controls can be stored at $\leq -70^{\circ}\mathrm{C}$ for 20 months with up to 2 freeze thaw cycles. Influenza A/Influenza B/RSV Reagent Mix can be frozen and thawed for up to 5 times, and M-MLV Reverse Transcriptase and RNase Inhibitor II up to 10 times, an increase from the current 5 freeze-thaw cycles. The ProFlu+ Assay package insert has been updated to reflect the current stability claims.
d. A declaration of conformity with design controls was submitted for the manufacturing facility which includes:
i) A statement signed by the Senior Director of R & D, Gen-Probe Prodesse, Inc., was submitted confirming that, as required by the risk analysis, all verification and validation
{5}
Page 6 of 6
activities were performed by the designated individual(s) and the results demonstrated that the predetermined acceptance criteria were met, and
ii) A “Declaration of Conformity” statement signed by the Associate Director of Quality and Regulatory, Gen-Probe Prodesse, Inc., was submitted stating that the manufacturing facility is in conformance with design control procedure requirements as specified in 21 CFR 820.30 and the records are available for review.
6. A Truthful and Accurate Statement, a 510(k) Summary or Statement and the Indications for Use Enclosure.
The labeling for this modified subject device has been reviewed to verify that the indication/intended use for the device is unaffected by the modification. In addition, the submitter's description of the particular modification(s) and the comparative information between the modified and unmodified devices demonstrate that the fundamental scientific technology has not changed. The submitter has provided the design control information as specified in The New 510(k) Paradigm and on this basis, I recommend the device be determined substantially equivalent to the previously cleared (or their preamendment) device.
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.