K211179 · Infervision Medical Technology Co., Ltd. · QAS · Aug 12, 2021 · Radiology
Device Facts
Record ID
K211179
Device Name
InferRead CT Stroke.AI
Applicant
Infervision Medical Technology Co., Ltd.
Product Code
QAS · Radiology
Decision Date
Aug 12, 2021
Decision
SESE
Submission Type
Traditional
Regulation
21 CFR 892.2080
Device Class
Class 2
Attributes
AI/ML, Software as a Medical Device, Real-World Evidence
Real-World Evidence
Submission
Device
Sponsor
RWD Sources
RWE Use Summary
Key Tags
K211179 · Aug 12, 2021
InferRead CT Stroke.AI
Infervision Medical Technology Co., Ltd.
Retrospective clinical CT scans from three U.S. hospitals
A retrospective study was conducted to evaluate the sensitivity and specificity of the InferRead CT Stroke.AI algorithm against a ground truth established by neuro-radiologists, and to compare the device's time-to-notification against standard-of-care time-to-open-exam metrics.
Retrospective study: 369 non-contrast brain CT scans from three U.S. hospitals
>1 (trained neuro-radiologists)
Indications for Use
InferRead CT Stroke.AI is a radiological computer aided triage and notification software for use in the analysis of Non-Enhanced Head CT images. The device is intended to assist hospital networks and trained radiologists in workflow triage by flagging suspected positive findings of intracranial hemorrhage (ICH). InferRead CT Stroke.AI uses an artificial intelligence algorithm to analyze images and highlight cases with detected ICH on a standalone desktop application in parallel to the ongoing standard of care image interpretation. The user is presented with a worklist with marked cases of suspected ICH findings. The device does not alter the original medical image, does not remove cases from queue, and is not intended to be used as a diagnostic device. If the clinician does not view the case, or if a case is not flagged, cases remain to be processed per the standard of care. The results of InferRead CT Stroke.AI are intended to be used in conjunction with other patient information and based on professional judgment, to assist with triage/prioritization of medical images. Notified clinicians are responsible for viewing full images per the standard of care.
Device Story
InferRead CT Stroke.AI is a radiological computer-aided triage and notification software. It receives non-enhanced head CT DICOM images; analyzes them using a deep learning algorithm to detect intracranial hemorrhage (ICH); and flags suspected positive cases on a desktop worklist. Used in hospital settings by radiologists, the device operates in parallel to standard-of-care image interpretation. It does not alter original images or remove cases from the queue. The radiologist views the flagged cases via a workstation viewer to prioritize clinical review. The device aims to reduce time-to-notification for potential ICH, facilitating faster clinical decision-making and triage.
Clinical Evidence
Retrospective study of 369 non-contrast head CT scans (51.5% positive for ICH, 48.5% negative) from three U.S. hospitals. Primary endpoints were sensitivity and specificity compared to neuro-radiologist ground truth. Results: sensitivity 0.916 (95% CI: 0.867-0.951), specificity 0.922 (95% CI: 0.872-0.957), and AUC 0.962. Time-to-notification (1.07 ± 0.57 min) was significantly faster than standard-of-care time-to-open-exam (75.4 ± 192.7 min, p < 0.001).
Technological Characteristics
Software-based radiological triage device; deep learning algorithm; runs on Ubuntu OS; deployed as an onsite server with client workstation interface. Comprises four modules: NeoViewer, Docking Toolbox, RePACS, and DLServer. Processes DICOM series; no hardware components; non-invasive.
Indications for Use
Indicated for use in the analysis of non-enhanced head CT images to assist hospital networks and trained radiologists in workflow triage by flagging suspected positive findings of intracranial hemorrhage (ICH).
Regulatory Classification
Identification
Radiological computer aided triage and notification software is an image processing prescription device intended to aid in prioritization and triage of radiological medical images. The device notifies a designated list of clinicians of the availability of time sensitive radiological medical images for review based on computer aided image analysis of those images performed by the device. The device does not mark, highlight, or direct users' attention to a specific location in the original image. The device does not remove cases from a reading queue. The device operates in parallel with the standard of care, which remains the default option for all cases.
Special Controls
Radiological computer aided triage and notification software must comply with the following special controls: 1. Design verification and validation must include: i. A detailed description of the notification and triage algorithms and all underlying image analysis algorithms including, but not limited to, a detailed description of the algorithm inputs and outputs, each major component or block, how the algorithm affects or relates to clinical practice or patient care, and any algorithm limitations. ii. A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide effective triage (e.g., improved time to review of prioritized images for pre-specified clinicians). iii. Results from performance testing that demonstrate that the device will provide effective triage. The performance assessment must be based on an appropriate measure to estimate the clinical effectiveness. The test dataset must contain sufficient numbers of cases from important cohorts (e.g., subsets defined by clinically relevant confounders, effect modifiers, associated diseases, and subsets defined by image acquisition characteristics) such that the performance estimates and confidence intervals for these individual subsets can be characterized with the device for the intended use population and imaging equipment. iv. Standalone performance testing protocols and results of the device. v. Appropriate software documentation (e.g., device hazard analysis; software requirements specification document; software design specification document; traceability analysis; description of verification and validation activities including system level test protocol, pass/fail criteria, and results). 2. Labeling must include the following: i. A detailed description of the patient population for which the device is indicated for use. ii. A detailed description of the intended user and user training that addresses appropriate use protocols for the device. iii. Discussion of warnings, precautions, and limitations must include situations in which the device may fail or may not operate at its expected performance level (e.g., poor image quality for certain subpopulations), as applicable. iv. A detailed description of compatible imaging hardware, imaging protocols, and requirements for input images. v. Device operating instructions. vi. A detailed summary of the performance testing, including: test methods, dataset characteristics, triage effectiveness (e.g., improved time to review of prioritized images for pre-specified clinicians), diagnostic accuracy of algorithms informing triage decision, and results with associated statistical uncertainty (e.g., confidence intervals), including a summary of subanalyses on case distributions stratified by relevant confounders, such as lesion and organ characteristics, disease stages, and imaging equipment.
*Classification.* Class II (special controls). The special controls for this device are:(1) Design verification and validation must include:
(i) A detailed description of the notification and triage algorithms and all underlying image analysis algorithms including, but not limited to, a detailed description of the algorithm inputs and outputs, each major component or block, how the algorithm affects or relates to clinical practice or patient care, and any algorithm limitations.
(ii) A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide effective triage (
*e.g.,* improved time to review of prioritized images for pre-specified clinicians).(iii) Results from performance testing that demonstrate that the device will provide effective triage. The performance assessment must be based on an appropriate measure to estimate the clinical effectiveness. The test dataset must contain sufficient numbers of cases from important cohorts (
*e.g.,* subsets defined by clinically relevant confounders, effect modifiers, associated diseases, and subsets defined by image acquisition characteristics) such that the performance estimates and confidence intervals for these individual subsets can be characterized with the device for the intended use population and imaging equipment.(iv) Stand-alone performance testing protocols and results of the device.
(v) Appropriate software documentation (
*e.g.,* device hazard analysis; software requirements specification document; software design specification document; traceability analysis; description of verification and validation activities including system level test protocol, pass/fail criteria, and results).(2) Labeling must include the following:
(i) A detailed description of the patient population for which the device is indicated for use;
(ii) A detailed description of the intended user and user training that addresses appropriate use protocols for the device;
(iii) Discussion of warnings, precautions, and limitations must include situations in which the device may fail or may not operate at its expected performance level (
*e.g.,* poor image quality for certain subpopulations), as applicable;(iv) A detailed description of compatible imaging hardware, imaging protocols, and requirements for input images;
(v) Device operating instructions; and
(vi) A detailed summary of the performance testing, including: test methods, dataset characteristics, triage effectiveness (
*e.g.,* improved time to review of prioritized images for pre-specified clinicians), diagnostic accuracy of algorithms informing triage decision, and results with associated statistical uncertainty (*e.g.,* confidence intervals), including a summary of subanalyses on case distributions stratified by relevant confounders, such as lesion and organ characteristics, disease stages, and imaging equipment.
{0}------------------------------------------------
Image /page/0/Picture/0 description: The image contains the logo of the U.S. Food and Drug Administration (FDA). On the left is the Department of Health & Human Services logo. To the right of that is the FDA logo, which is a blue square with the letters "FDA" in white. To the right of the blue square is the text "U.S. FOOD & DRUG ADMINISTRATION" in blue.
August 12, 2021
Infervision Medical Technology Co., Ltd. % Mr. Matt Deng Director Infervision US, Inc. 1900 Market Street PHILADELPHIA PA 19103
Re: K211179
Trade/Device Name: InferRead CT Stroke.AI Regulation Number: 21 CFR 892.2080 Regulation Name: Radiological computer aided triage and notification software Regulatory Class: Class II Product Code: QAS Dated: July 8, 2021 Received: July 12, 2021
Dear Mr. Deng:
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 requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part 801 and Part 809); medical device reporting (reporting of medical device-related adverse events) (21 CFR
{1}------------------------------------------------
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 (QS) 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.
For
Thalia T. Mills, Ph.D. Director Division of Radiological Health OHT7: Office of In Vitro Diagnostics and Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health
Enclosure
{2}------------------------------------------------
# Indications for Use
510(k) Number (if known) K211179
Device Name InferRead CT Stroke.AI
#### Indications for Use (Describe)
InferRead CT Stroke.AI is a radiological computer aided triage and notification software for use in the analysis of Non-Enhanced Head CT images. The device is intended to assist hospital networks and trained radiologists in workflow trage by flagging suspected positive findings of intracranial hemorrhage (ICH).
InferRead CT Stroke.Al uses an artificial intelligence algorithm to analyze images and highlight cases with detected ICH on a standalone desktop application in parallel to the ongoing standard of care image interpretation. The user is presented with a worklist with marked cases of suspected ICH findings. The device does not alter the original medical image, does not remove cases from queue, and is not intended to be used as a diagnostic device. If the clinician does not view the case, or if a case is not flagged, cases remain to be processed per the standard of care.
The results of InferRead CT Stroke.AI are intended to be used in conjunction with other patient information and based on professional judgment, to assist with triage/prioritization of medical images. Notified clinicians are responsible for viewing full images per the standard of care.
| Type of Use (Select one or both, as applicable) | |
|-------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------|
| <div><span style="font-family: Arial, sans-serif;">\[X]</span> Prescription Use (Part 21 CFR 801 Subpart D)</div> | <div><span style="font-family: Arial, sans-serif;">[]</span> Over-The-Counter Use (21 CFR 801 Subpart C)</div> |
#### 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
## Infervision Medical Technology Co., Ltd. K211179
This 510(k) Summary is in conformance with 21 CFR 807.92
| Submitter: | Infervision Medical Technology Co., Ltd.<br>Room B401, 4th Floor, Building 1,<br>No. 12 Shangdi Information Road,<br>Haidian District, Beijing, 100085, China<br>Phone: +86 10-86462323 |
|--------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Primary Contact: | Matt Deng<br>Email: matt.deng@infervision.ai<br>Phone: 919-886-6082 |
| Second Primary<br>Contact: | Frank Wu<br>Email: frank.wu@infervision.ai<br>Phone: 857-988-1888 |
| Company Contact: | Xiaoyan Fan<br>Email: fxiaoyan@infervision.com<br>Phone: +86 13810508664 |
| Date Prepared: | April 8, 2021 |
| Device Name and Classification | |
| Trade Name: | InferRead CT Stroke.AI |
| Common Name: | Radiological computer aided triage and notification software |
| Classification: | Class II |
21 CFR 892.2080, Radiological computer aided triage and Regulation Number: notification software Classification Panel: Radiology Product Code:
QAS
## Predicate Device:
| Primary Predicate | |
|-------------------------|-----------------|
| Trade Name | BriefCase |
| 510(k) Submitter/Holder | Aidoc |
| Class | Class II |
| Regulation Number | 21 CFR 892.2080 |
| Classification Panel | Radiology |
| Product Code | QAS |
{4}------------------------------------------------
### Device Description
InferRead CT Stroke.AI is a radiological computer-assisted triage and notification software device. The software device is a computer program with a deep learning algorithm running on Ubuntu operating system. The device can be deployed as an onsite server in the hospital and the user interacts with the software from a client workstation. The device can be broken down into 4 modules, the NeoViewer, Docking Toolbox, RePACS, and DLServer.
The Docking Toolbox module receives DICOM series and inspects the series against a list of requirements. Series that pass the requirements are sent into the system for prediction for intracranial hemorrhage. Series are processed in a first-out order. When hemorrhage is detected, the system marks the case in the work list prompting the user to conduct preemptive triage and prioritization.
When the user refreshes the page, cases with suspected findings will be marked with an indicator. Cases are identified, such as by Name and Patient ID. The user may filter and sort by suspected ICH and identify the case. A preview is available but is not intended for primary diagnosis and a radiologist must review the case per their standard process. The suspected cases assist in triaging intracranial hemorrhage cases sooner than standard of care practice alone.
#### Intended Use/Indications for Use
InferRead CT Stroke.AI is a radiological computer aided triage and notification software for use in the analysis of non-enhanced head CT images. The device is intended to assist hospital networks and trained radiologists in workflow triage by flagging suspected positive findings of intracranial hemorrhage (ICH).
InferRead CT Stroke.AI uses an artificial intelligence algorithm to analyze images and highlight cases with detected ICH on a standalone desktop application in parallel to the ongoing standard of care image interpretation. The user is presented with a worklist with marked cases of suspected ICH findings. The device does not alter the original medical image, does not remove cases from queue, and is not intended to be used as a diagnostic device. If the clinician does not view the case, or if a case is not flagged, cases remain to be processed per the standard of care.
The results of InferRead CT Stroke.AI are intended to be used in conjunction with other patient information and based on professional judgment, to assist with triage/prioritization of medical images. Notified clinicians are responsible for viewing full images per the standard of care.
{5}------------------------------------------------
### Comparison of Technological Characteristics
The subject and predicate devices are radiological computer-assisted triage and notification software. Both devices are artificial intelligence algorithms incorporated software packages for use with CT scanners, PACS, and workstations. Both devices process images intended to aid in prioritization and triage of non-enhanced head CT cases with intracranial hemorrhage. Both have the same intended use and indications for use for flagging suspected cases, and indicating to the clinician for review.
The predicate device sends pop up notifications and compressed previews to the workstations of radiologist. The subject device doesn't send pop up notifications as the predicate device. Instead, to fulfill the notification, the subject device visually marks the case in the worklist, indicating to a radiologist the need to review those images for ICH. For both devices, the user must be alert and receptive to the outputs of the device. Similarly to predicate device, the subject device also works in parallel to the standard of care. The indication prompts preemptive triage of the flagged case where the radiologist may decide to perform evaluation. Similarly, if the notification is rejected, the case remains in their standard queue to be handled per their standard of care.
The subject device provides a viewer on the workstation allowing the radiologist to preview the DICOM similarly to the compressed preview of the subject device. This viewer allows the user to Scroll through series. Similar to the predicate, the preview is for informational purposes only and not for diagnostic use. The notified clinicians are responsible for using the local imaging system for viewing the original images and engage the referring clinician for diagnosis and treatment decisions.
{6}------------------------------------------------
The subject and predicate software utilizes a deep learning algorithm trained on medical images. The same type of safety and effectiveness questions as the predicate. That is, accurate detection of intracranial hemorrhage within the study on which a physician can base a clinically useful triage/prioritization assessment considering all available clinical information. Like the predicate, the subject device does not reading queue. Both devices operate in parallel with the standard of care, which remains the default option for all cases.
| Item | InferRead CT Stroke.AI<br>(Subject Device) | Aidoc Briefcase ICH (K180647)<br>(Predicate Device) | Comparison |
|--------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Intended Use /<br>Indications for Use | InferRead CT Stroke.AI is a<br>radiological computer aided triage<br>and notification software for use in<br>the analysis of Non-Enhanced Head<br>CT images. The device is intended to<br>assist hospital networks and trained<br>radiologists in workflow triage by<br>flagging suspected positive findings<br>of intracranial hemorrhage (ICH).<br><br>InferRead CT Stroke.AI uses an<br>artificial intelligence algorithm to<br>analyze images and highlight cases<br>with detected ICH on a standalone<br>desktop application in parallel to the<br>ongoing standard of care image<br>interpretation. The user is presented<br>with a worklist with marked cases of<br>suspected ICH findings. The device<br>does not alter the original medical | BriefCase is a radiological computer<br>aided triage and notification software<br>indicated for use in the analysis of non-<br>enhanced head CT images. The device is<br>intended to assist hospital networks and<br>trained radiologists in workflow triage<br>by flagging and communication of<br>suspected positive findings of<br>pathologies in head CT images, namely<br>Intracranial Hemorrhage (ICH).<br><br>BriefCase uses an artificial intelligence<br>algorithm to analyze images and<br>highlight cases with<br>detected ICH on a standalone desktop<br>application in parallel to the ongoing<br>standard of care<br>image interpretation. The user is<br>presented with notifications for cases<br>with suspected ICH | InferRead CT Stroke.AI<br>and the previously cleared<br>BriefCase (K180647)<br>have the same intended<br>use and\indications for use<br>in terms of finding<br>suspected intracranial<br>hemorrhage in non<br>contrast head CT, flagging<br>suspected cases, and<br>indicating the case to the<br>attention of the clinician. |
| | image, does not remove cases from queue, and is not intended to be used as a diagnostic device. If the clinician does not view the case, or if a case is not flagged, cases remain to be processed per the standard of care.<br><br>The results of InferRead CT Stroke.AI are intended to be used in conjunction with other patient information and based on professional judgment, to assist with triage/prioritization of medical images. Notified clinicians are responsible for viewing full images per the standard of care. | findings. Notifications include compressed preview images that are meant for informational purposes only and not intended for diagnostic use beyond notification. The device does not alter the original medical image and is not intended to be used as a diagnostic device.<br><br>The results of BriefCase are intended to be used in conjunction with other patient information and based on professional judgment, to assist with triage/prioritization of medical images. Notified clinicians are responsible for viewing full images per the standard of care. | Both are designed to be used by the radiologist, prompt the radiologist to start preemptive triage of a flagged case |
| User Population | Radiologist | Radiologist | Both are indicated for use in analysis of non- enhanced head CT |
| Anatomical Region of Interest | Head | Head | |
| Data Acquisition Protocol | Non contrast CT scan of the head | Non contrast CT scan of the head or neck | |
| View DICOM data | DICOM information about the patient, study and current image | DICOM information about the patient, study and current image | Both display DICOM information for informational purposes only |
| | | | |
| Segmentation of<br>region of interest | No; device does not mark, highlight,<br>or direct users' attention to a specific<br>location in the original image | No; device does not mark, highlight, or<br>direct users' attention to a specific<br>location in the original image | Neither marks, highlights<br>or directs attention to a<br>specific location in the<br>original image |
| Algorithm | Artificial intelligence algorithm with<br>database of images | Artificial intelligence algorithm with<br>database of images | Both use artificial<br>intelligence algorithm<br>with a database of images |
| Notification /<br>Prioritization | Yes, Case level indicator | Yes, pop-up notifications, case level<br>indicator | In both, the suspected<br>cases are indicated to the<br>user. The subject device<br>provides case level<br>indicator and allow the<br>user to sort suspected<br>cases to the top. |
| Preview Images | Presentation of a preview of the<br>study for initial assessment not<br>meant for diagnostic purposes<br>The device operates in parallel<br>with the standard of care, which<br>remains the default option for all<br>cases | Presentation of a preview of the<br>study for initial assessment not<br>meant for diagnostic purposes<br>The device operates in parallel<br>with the standard of care, which<br>remains the default option for all<br>cases | Both allow the user to<br>view the image. The<br>device is intended to work<br>in parallel with standard<br>of care. |
| Alteration of<br>original image | No | No | Neither alters the original<br>image. |
| Removal of cases<br>from worklist<br>queue | No | No | Neither removes cases<br>from the worklist queue. |
Detailed Comparison of the Subject and Predicate Devices
{7}------------------------------------------------
{8}------------------------------------------------
{9}------------------------------------------------
### Performance Data
Infervision conducted a retrospective study to assess the clinical performance and notification functionality of the InferRead CT Stroke.AI software. The study evaluated the InferRead deep learning algorithm in terms of sensitivity and specificity with respect to a ground truth, as established by trained neuro-radiologists, in the detection of intracranial hemorrhage (ICH) in the brain. In addition, the study reported and compared the InferRead time-to-notification and the Time-to-open-exam for the standard of care. The InferRead time-to-notification includes the time to receive the DICOM scan, analyze and the worklist application shows the results. The standard of care time-to-open-exam consisted of the time from the initial scan of the patient to when the radiologist first opens the exam for review.
A total of 369 non-contrast brain CT scans (studies) were obtained from three hospitals in the U.S. There were approximately equal numbers of positive and negative cases (51.5% of images with ICH and 48.5% without ICH, respectively) included in the analysis. Comparing the InferRead software output to the ground truth, the sensitivity and specificity of InferRead CT Stroke.AI are 0.916 (95% CI: 0.867-0.951) and 0.922 (95% CI: 0.872-0.957), which are significantly higher than the 80% null hypothesis (p values < 0.001). This study met the pre-specified performance goals of 80% for sensitivity and specificity.
In addition, the area under the receiver operating characteristic curve (AUC) was 0.962, demonstrating the clinical utility and potential benefits of the InferRead software based on the imaging study results.
The InferRead time-to-notification is 1.07 ± 0.57 (mean ± SD) minutes, which is substantially lower than the standard of care time-to-open-exam of 75.4±192.7 minutes (P < 0.001). This validation study shows that InferRead CT Stroke.AI is both safe and effective.
#### Conclusions
Based on the technological comparison, the major difference is that the subject device marks the suspected case in the work list instead of a pop up notification. In both cases, the user must be receptive to the visual outputs of the device. After comparison, we do not find different issues of safety or effectiveness. In the same way as the predicate, the subject also uses deep learning algorithm to process images and predict intracranial hemorrhage. The intended use of the two are the same.
The performance testing demonstrated that the subject device performed similarly. These results show that the subject device does not have significant differences and is as safe and effective as a legally marketed predicate 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.
