K980213 · Lumisys, Inc. · LMA · Mar 17, 1998 · Radiology
Device Facts
Record ID
K980213
Device Name
DI-2000
Applicant
Lumisys, Inc.
Product Code
LMA · Radiology
Decision Date
Mar 17, 1998
Decision
SESE
Submission Type
Traditional
Regulation
21 CFR 892.2030
Device Class
Class 2
Attributes
Software as a Medical Device
Indications for Use
DI-2000 is intended to utilize a scanner and software interface to digitize either radiology film or computed radiography exposed phosphor plates. DI-2000 is a DICOM 3.0 compliant radiological digitization application. DI-2000 enables the user to autoarchive lossless or lossy compressed images locally or at a remote archive site. Supports DICOM 3.0 Query and Retrieve Service Class. Supports scanning of films or phosphor plates in batch mode prior to entering patient information. Once scanned, images can be sent to multiple destinations. Users: The DI-2000 will be located in a radiology department or clinic, or in a mobile radiology van. DI-2000 will support reading and digitizing exposed CR plates or radiology film. The expected users are as follows: Radiologists Radiology Technologists System Administrator within the radiology department
Device Story
DI-2000 is a DICOM 3.0-compliant software application for digitizing radiology film and computed radiography (CR) phosphor plates. It interfaces with Lumisys digitizers to capture images, which are then processed, compressed (JPEG or wavelet), and managed via a local Sybase SQL database. Operated by radiologists, technologists, or administrators in clinical settings, the device supports batch scanning, local/remote archiving, and network transmission. It provides image manipulation tools (flip, rotate, inversion) and quality assurance functions. The system uses automated checks to prevent user errors, such as duplicate IDs or unsaved images. By enabling digital conversion and standardized DICOM communication, the device facilitates integration of analog or CR images into PACS environments, supporting clinical decision-making through improved image accessibility and archival efficiency.
Clinical Evidence
No clinical data. Bench testing only. Software hazard analysis and functional validation were performed to ensure system reliability, image quality maintenance during compression, and adherence to DICOM standards.
Technological Characteristics
Software-based DICOM 3.0 application; runs on Windows NT 4.0; utilizes Sybase SQL Anywhere database. Supports JPEG and wavelet image compression. Connectivity via 10/100 Base-T Ethernet, ATM, ISDN, and modems. Hardware requirements: Pentium 200MHz, 64MB RAM, 2GB HDD. Complies with ACR/NEMA DICOM 3.0, CCITT, and ISO/IEC 10918-1 standards.
Indications for Use
Indicated for use by radiologists, radiology technologists, and system administrators in radiology departments, clinics, or mobile vans to digitize radiology film or computed radiography phosphor plates for DICOM-compliant archiving, retrieval, and transmission.
Regulatory Classification
Identification
A medical image digitizer is a device intended to convert an analog medical image into a digital format. Examples include systems employing video frame grabbers, and scanners which use lasers or charge-coupled devices.
Special Controls
*Classification.* Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std.). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 892.9.
{0}------------------------------------------------
K98 0213
# Addendum to 510(k) for DICOM Client K955708
## 510(k) Summary
#### 1. Identification
MAR 1 7 1998
January 7, 1998 Date:
Lumisys, Inc. Submitter: 1350 North Kolb Tucson, AZ 85715
Contact:
Trindy LeForge (520) 751-6842 (520) 298-1400 (FAX)
### Device Name 2.
Trade Name: DI-2000 DICOM Client Common Name:
### Registration Number 3.
2029024
### 4. Classification
No formal classifications have been issued for PACS components. For purposes of determining substantial equivalence they have been considered to be accessories to medical imaging devices. A stationary x-ray system is a permanently installed diagnostic system intended to generate and control x-rays for examination of various anatomical regions. Accessories to a stationary x-ray system, such as digitizing application software, have been classified as Class II Radiology Devices under 21 CFR 892.1680.
#### 5. Standards
Performance: Voluntary:
None established. ACR/NEMA DICOM 3.0
Requirements of 21 CFR 1020.10.
CCITT (Consultative Committee on International Telephone and Telegraph) and ISO/IEC 10918-1 for compression.
### 6. Labeling
See Appendix A.
{1}------------------------------------------------
### Safety and Effectiveness Information 7.
### Hazard Analysis
.
.
Formal analysis of the design for potential and existing hazards takes place during the requirements phase, the design phase, and during the testing phase. All potential existing hazards are analyzed using the following steps:
- Determine if hazard exists in the product. Hazard Identification .
- Level of Concern Determine whether the hazard is of minor, moderate, or major concern.
- Determine all possible causes through analysis or testing. Possible Causes .
- Identify possible methods of controlling or eliminating the Methods of Control hazard.
- Decide upon corrective action to mitigate the hazard. Corrective Action .
The DI-2000 software was analyzed for potential hazards. A summary of hazards that were considered and found not to be exposures in the design of DI-2000 are included in Appendix E.
### Level of Concern
The level of concern for the DI-2000 software is considered low for the following reasons:
- If the transfer of data cannot be completed without compromising the image quality, the 1. transaction cannot be completed, and therefore, the image quality cannot be lost.
- The system provides checks and balances for potential user errors, such as failing to save 2. images, or assigning duplicate IDs. The system will not allow users to proceed until images are cleared or saved appropriately.
- These checks and balances are re-validated with each release of the product. Separate test 3. plans exercise these specific points within the system. Products are not released until all checks and balances are fully validated.
#### 8. Photographs
See Appendix B.
#### 9. Device Description
A detailed description of each of all functions is contained in the Functional Requirements Specification, included as Appendix B. The following feature comparison provides an easy way to identify the changes between the initial 510(k) for DICOM Client submitted December 1995, and this addendum for DI-2000.
| Features | DICOM Client | DI-2000 |
|-------------------------------------------------|-----------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Patient Information Modification | Y | Y |
| Delete Patient Information | Y | Y |
| Fax Report | Y | N |
| Print DICOM Image | Y | Y |
| Export DICOM images to TIFF | Y | N |
| Annotate DICOM Images | Y | N |
| Window/Level Preset Settings | Y | N |
| Image Flip | Y | Y |
| 90 degree Image Rotation | Y | Y |
| Features | DICOM Client | DI-2000 |
| 180 degree Image Rotation | Y | Y |
| Black/White Inversion | Y | Y |
| Multiple Image Display | Y | N |
| Token/Iconic Display | Y | Y |
| DICOM Send | Y | Y |
| DICOM Receive | Y | Y |
| DICOM Query User | Y | N |
| DICOM Query Provider | Y | Y |
| DICOM Retrieve User | Y | N |
| DICOM Retrieve Provider | Y | Y |
| DICOM Echo | Y | Y |
| LAN/WAN Communication | Y | Y |
| Zoom/Unzoom | Y | N |
| Copy to Clipboard | Y | N |
| JPEG Compression lossy/lossless | Y | Y |
| iJPEG Compression | Y | Y |
| Analog modem support | Y | Y |
| Lumisys Film Digitizers connection capability | Y | Y |
| Industry Standard Digital Communication Support | Y | Y |
| Display Resolution | Up to 2K | Up to 2K |
| Communication Media Supported | ATM, ISDN,<br>Analog modems,<br>Ethernet, FDDI,<br>Token Ring,<br>Analog Phone<br>Lines | ADSL, Cable<br>Modem, Routers,<br>Brouters, Internet,<br>ATM, ISDN,<br>Analog modems,<br>Fast Ethernet,<br>Gigabit Ethernet,<br>Ethernet, FDDI,<br>Token Ring, Analog<br>Phone Lines |
| DICOM Print | N | Y |
| Images Displayed | CT, MR, US, CR,<br>NM, Plain Films | CT, MR, US, CR,<br>NM, Plain Films |
| Color Images | Y | Y |
| Cine Loop Viewing | Y | Y |
| Measurement Tools | Y | Y |
| DICOM Removable Media Support | N | Y |
| Wavelet Compression | N | Y |
{2}------------------------------------------------
## DI-2000 Components
The following is a list of the minimum recommended requirements for the DI-2000 software:
- Pentium 200MHz CPU or equivalent. 1.
- 64MB of RAM 2.
- 2MB video board capable of 1024-by-760 screen resolution at 16K colors. 3.
- 14-inch monitor capable of displaying a 1024-by-760 screen. 4.
- న. 2GB local hard drive
- Iomega Corp JAZ drive for local archive and redundancy capabilities. 6.
- Network connection of 10/100 Base-T Ethernet card. 7.
{3}------------------------------------------------
- Microsoft Windows NT Workstation 3.51 or 4.0 operating system with Service Pack 3 or 8. later.
- Sybase SQL Anywhere installed as the local database. 9.
- 10. Symantec Corp. pcANYWHERE version 8.0.
### Comparison with Equivalent Devices 10.
Lumisys, Inc., believes that DI-2000 is substantially equivalent to the following medical devices in commercial distribution:
| Manufacturer | Product (Trade Name) | 510(k)# | Decision Date |
|--------------------------------|----------------------------------------------------------------------------------------------------------------|---------|---------------|
| Howtek, Inc. | Digitizer Director™ | K972191 | 9/8/97 |
| DeJarnette Research<br>Systems | ImageShare Film Digitizer 1000™ | K963628 | 11/7/96 |
| DeJarnette Research<br>Systems | ImageShare Secondary Capture<br>Acquisition Station; ImageShare<br>Computed Radiography Acquisition<br>Station | K963594 | 11/7/96 |
| Mitra Imaging, Inc. | Film Express | K970652 | 5/2/97 |
### Feature Comparison
| Feature | DI-2000 | Digitizer<br>Director™ | ImageShare™ | Film Express |
|----------------------------------------------------------------|---------|------------------------|-------------|--------------|
| Bulk Transfer | Y | N | N | N |
| DICOM 3.0 Compliance | Y | Y | Y | Y |
| Windows NT 4.0 | Y | Y | N | Y |
| Internet Connectivity | Y | Y | Y | N |
| Variable Compression | Y | N | Y | N |
| Wavelet Compression | Y | N | N | N |
| JPEG Lossless, Lossy, Enhanced<br>Compression | Y | N | N | Y |
| Remote Administration and<br>Management Support | Y | N | N | N |
| Image Processing Algorithm | Y | N | MANUAL | N |
| Automatic Disk Maintenance | Y | N | N | N |
| Auto-Transmit | Y | N | N | Y |
| Supports Lumisys, Inc., digitizers<br>and Computed Radiography | Y | N | Y | Y |
| Quality Assurance Function | Y | Y | Y | Y |
| Patient Identification on the Plate | Y | N | N | N |
| Local Archive Functionality | Y | Y | N | N |
| Batch Mode Scanning | Y | Y | N | N |
| Network Printing Support | Y | N | Y | N |
| Confirmed Delivery | Y | N | N | Y |
### 11. Image Compression Technology
DI-2000 supports the following compression methodologies:
{4}------------------------------------------------
| Compression Type | Algorithms |
|------------------|--------------------------|
| Lossless | In hardware, via network |
| | JPEG |
| | Wavelet |
| Lossy | JPEG |
| | Wavelet |
In general, the software conforms to the image formats described in the DICOM standard.
### Wavelet compression description.
DI-2000 supports the use of wavelet compression in addition to JPEG (described in the original submission of the 510(k), K955708, December 5, 1995.) Images that have been compressed using the wavelet technique are included in Appendix B.
Wavelet transform based compression involves the representation of an image at a variety of scales of resolution. The wavelet transform itself is performed by repeatedly filtering the image with a pair of high and low pass digital filters which obey strict mathematical constraints. Often referred to as a multi-resolution decomposition, the wavelet transform allows fine details in an image (otherwise known as "contrast" information) to be separated from larger scale trends. In this implementation, the requested compression ratio controls the degree of contrast identified. After an image has undergone a wavelet transform, an effort is made to detect those regions of the transform which have little or no contrast. Once such a region has been identified, it is quantized (stored with fewer bits of precision than those parts of the transform which appear important). It is in the quantization steps that all loss occurs. Once quantization is performed, it is not possible to retrieve the original, higher presentation. Following quantization, standard lossless compression techniques are used which includes run-length encoding. These techniques perform quite well as they are generally operating on data which contains large runs of zero with no appreciable effect on image quality.
As the rows of the image are sent to the compressor, the lengths are padded to a multiple of 16 and then the row transform is performed. After all of the lines have been sent, the last line is copied as many times as needed to pad the height to a multiple of 16. The image is then sent off to have its columns transformed; this completes the first-level decomposition
| HH | HL |
|----|----|
| LH | HH |
Figure 1 Subband structure of single-level wavelet transform
For each of the three high-frequency quadrants, an energy mapping is computed. Each quadrant is broken up into 8x8 sub-blocks and the energy of each sub-block is computed.
{5}------------------------------------------------
These energy values are sent off to a clustering algorithm which partitions them into 4 levels. Each 8x8 sub-block in the quadrant is labeled with one of the four levels. The resulting three mappings are stored to be used later in partitioning of the subbands.
| 1 | 1 | 2 | 4 |
|---|---|---|---|
| 3 | 3 | 2 | 2 |
| 1 | 4 | 3 | 3 |
| 1 | 2 | 2 | 1 |
Figure 2 Example of energy partitioning map of a high-frequency quadrant.
The decomposition is finished by doing one more level in the high frequency quadrants and then performing the standard decomposition in the LL quadrant until the lowest resolution subband image dimensions are at some specified level.
Table 3 Subband structure of a full wavelet decomposition.
After the decomposition, the 2nd, 3rd, and 4th level subbands are each partitioned into 4 sequences by the mapping constructed in step 2. The 2nd level subbands are subdivided into 4x4 subblocks and one of the mappings is superposed on this subdivision. The 3d level subbands are subdivided into 2x2 subblocks and, again, one of the mappings are superposed on this subdivision. Finally, one of the mappings is superposed onto the 4" level subbands. Recall that in step 2, three mappings were constructed. The particular mappings applied to each of the 2nd-4th level subbands can be obtained by referring to figure 3. These mappings partition each of the 2nd-4th level subbands into 4 sequences which constitute one part of the bands making up the transformed data Each of the remaining higher level subbands form the other set of transform bands. For each transform band, we calculate the normalized kurtosis, the 40 moment divided by the squared variance, and then matched to a generalized Gaussian distribution. The normalized kurtosis of a generalized Gaussian distribution with exponent $\alpha$ and variance $\sigma$2 is given by
{6}------------------------------------------------
$$\frac{E\left[\mathbf{x}^{*}\right]}{\sigma^{2}} = \frac{\Gamma(\frac{1}{\alpha})}{\Gamma(\frac{3}{\alpha})\Gamma(\frac{1}{\alpha})} \frac{1}{\alpha}.$$
where I is the ordinary Gamma function. The kurtosis is then matched to one of the predetermined generalized Gaussian exponents .5, .75, 1.0, 1.5, 2.0. This exponent, along with the variance, is used to supply the generalized Gaussian pdf model for the given transform band.
The transform bands, along with there model parameters a and o2 and the length of the band, are sent to a bit-allocator which then returns the step sizes to be used in the quantization of each transform band.
### 12. Test Data and Conclusions
See Appendix B.
### 13. Software Information
For a list and brief description of software functions, a description of the software development methods, and a description of software test procedures, see Appendix C.
### Truthful and Accurate Statement 14.
As required by 21 CFR 807.87(j) a Truthful and Accurate Statement is contained in Appendix D.
### 15. Confidentiality
Portions of the attached enclosure are confidential.
{7}------------------------------------------------
# DEPARTMENT OF HEALTH & HUMAN SERVICES
Public Health Service
Food and Drug Administration 9200 Corporate Boulevard Rockville MD 20850
Trindy LeForge Director, Ouality Assurance Lumisys, Inc. 1350 North Kalb Road Tucson, Arizona 85715
Re:
K980213 DI-2000 (Digital Interface) Dated: January 7, 1998 Received: January 9, 1998 Regulatory class: Unclassified Procode: 90 LMA
MAR 17 1998
Dear Mr. LeForge:
We have reviewed your Section 510(k) notification of intent to market the device referenced above and we have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to devices marketed in interstate commerce prior to May 28, 1976, the castment 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.). You may, therefore, market the device, subject to the general controls provisions of the Act. 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.
If your device is classified (see above) into either class II (Special Controls) or class III (Premarket Approval), it may be subject to such additional controls. Existing major regulations affecting your device can be found in the Code of Feberal Regulations. Title 21, Parts 800 to 895. A substantially equivalent determination assumes compliance with the Current Good Manufacturing Practice requirements, as set forth in the Quality System Regulation (OS) for Medical Devices: General regulation (21 CFR Part 820) and that, through periodic QS inspections, the Food and Drug Administration (FDA) will verify such assumptions. Failure to comply with the GMP regulation may result in regulatory action. In addition, FDA may publish further announcements concerning your device in the Federal Register. Please note: this response to your premarket notification submission does not affect any obligation you might have under sections 531 through 542 of the Act for devices under the Electronic Product Radiation Control provisions, or other Federal laws or regulations.
This letter will allow you to begin marketing your device as described in your 510(k) premarket notification. The FDA finding of substantial equivalence of your device to a legally marketed predicate device results in a classification for your device and thus, permits your device to proceed to the market.
If you desire specific advice for your device on our labeling regulation (21 CFR Part 801 and additionally 809.10 for yitro diagnostic devices), please contact the Office of Compliance at (301) 594-4613. Additionally, for questions on the promotion and advertising of your device, please contact the Office of Compliance at (301) 594-4639. Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR 807.97). Other general information on your responsibilities under the Act may be obtained from the Division of Small Manufacturers Assistance at its toll-free number (800) 638-2041 or (301) 443-6597 or at its Internet address "http://www.fda.gov/cdrb/dsma/dsmamain.httml":
Sincerely yours,
Lillian Xia, Ph.D.
Lillian Yin, Ph.D. Director, Division of Reproductive Abdominal, Ear, Nose and Throa and Radiological Devices Office of Device Evaluation Center for Devices and Radiological Health
Enclosure
Image /page/7/Picture/15 description: The image shows a partial view of a logo or emblem, featuring the text 'DEPARTMENT OF HEALTH & HU'. The text is arranged in a curved manner, suggesting it is part of a circular design. The visible portion of the emblem includes stylized graphic elements, possibly representing abstract shapes or symbols associated with health or the department it represents.
{8}------------------------------------------------
4955708- 1498 0213 510(k) Number:
DICOM Client Device Name: DI-2000
### Indications for Use:
DI-2000 is intended to utilize a scanner and software interface to digitize either radiology film or computed radiography exposed phosphor plates.
DI-2000 is a DICOM 3.0 compliant radiological digitization application.
DI-2000 enables the user to autoarchive lossless or lossy compressed images locally or at a remote archive site. Supports DICOM 3.0 Query and Retrieve Service Class.
Supports scanning of films or phosphor plates in batch mode prior to entering patient information. Once scanned, images can be sent to multiple destinations.
Users:
The DI-2000 will be located in a radiology department or clinic, or in a mobile radiology van. DI-2000 will support reading and digitizing exposed CR plates or radiology film. The expected users are as follows:
Radiologists Radiology Technologists System Administrator within the radiology department
Thiel le. S. ymm
(Division Sign-Off) Division of Reproductive, Abdominal, EN and Radiological Devic 510(k) Number
**Prescription Use**
(Per 21 CFR 801.109)
Lumisys, Inc.
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.