(230 days)
Heuron ICH is radiological computer-aided triage and notification software designed for the analysis of non-contrast head CT images in adults or transitional adolescents aged 18 and older. This device is intended to aid appropriately trained medical specialists and hospital networks in streamlining workflow by identifying and communicating suspected positive findings of Intracranial hemorrhage (ICH).
Heuron ICH employs an artificial intelligence algorithm to analyze non-contrast CT images, flagging cases with identified findings through a dedicated application that operates in parallel with the standard of care image interpretation process. Users receive notifications for cases with suspected findings, which include compressed preview images provided for informational purposes only and are not intended for diagnostic use beyond notification. Importantly, Heuron ICH does not modify the original medical images and is not intended to serve as a diagnostic device.
The results generated by Heuron ICH are intended to complement other patient information and assist medical specialists in prioritizing and triaging medical images. Notified medical specialists are responsible for viewing the full non-contrast CT images in accordance with established standard of care practices.
Heuron ICH registers with the hospital's Picture Archiving and Communication System (PACS) using IP, Port, AE title, and TLS authentication details. It automatically receives Non-Contrast Computed Tomography (NCCT) images in DICOM format from PACS. Upon connection request from PACS to Heuron ICH, the system verifies the IP, Port, AE title, and TLS authentication information before accepting the image transmission. The product does not query PACS to retrieve images. Instead, it receives images automatically from PACS systems that are allowed access by registering a list (white list) of PACS systems capable of uploading images to the product.
The Heuron ICH is an artificial intelligence-based solution that analyzes non-contrast CT images and provides a notification of suspected positive cases of intracranial hemorrhage (ICH) for prioritization of review. Heuron ICH uses deep learning (DL) technique of a convolutional neural network (CNN). Dataset obtained from the RSNA (Radiological Society of North America) Brain CT Hemorrhage Challenge 2019 was used for training and development of the model. Once the DICOM images transmitted from PACS are uploaded to the Heuron ICH server, the images become accessible through the worklist. The worklist displays patient identification information (Patient ID, name, age, etc.) and analysis status for convenient reference. Images received by Heuron ICH server are analyzed in the order of reception.
During the analysis, if ICH is suspected, the server provides users with a notification. The notifications include compressed preview images, which are not to be used for diagnostic use, but only for informational purposes. It is important to note that the software does not provide segmentation, analysis, or intermediate outputs to users. These notifications can be sent to registered email addresses, mobile SMS, and through the mobile app push notification feature.
Here's a breakdown of the acceptance criteria and the study proving the Heuron ICH device meets these criteria, based on the provided FDA 510(k) Clearance Letter.
Acceptance Criteria and Device Performance
1. Table of Acceptance Criteria and Reported Device Performance
| Metric | Acceptance Criteria (Lower Bound of 95% CI) | Reported Device Performance (Value and 95% CI) | Met Criteria? |
|---|---|---|---|
| Sensitivity | 80% | 86.3% (95% CI: 81.9-90.3) | Yes |
| Specificity | 80% | 87.6% (95% CI: 83.9-91.0) | Yes |
Note: The document specifies the acceptance criteria as the lower bound of the 95% Confidence Interval for both sensitivity and specificity.
Study Details
2. Sample Size and Data Provenance
- Test Set Sample Size: 600 NCCT images
- Data Provenance:
- Country of Origin: United States (obtained from three different hospitals located in the US)
- Retrospective or Prospective: Retrospective
3. Number of Experts and Qualifications for Ground Truth
- Number of Experts: 3
- Qualifications of Experts: US board-certified neuroradiologists
4. Adjudication Method for the Test Set
- Adjudication Method: 2+1. Two US board-certified neuroradiologists (truthers) independently interpreted each NCCT image. In case of disagreement between these two, a third truther reviewed the case to establish the final ground truth.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- Was an MRMC study done? No. The document describes a "standalone performance study." While it mentions "time-to-notification" comparison to standard of care, it doesn't detail a comparative effectiveness study involving human readers with and without AI assistance for diagnostic accuracy improvements.
6. Standalone Performance Study
- Was a standalone (algorithm only) performance study done? Yes. The document explicitly states, "The standalone performance study results exceeded the acceptance criteria..."
7. Type of Ground Truth Used
- Type of Ground Truth: Expert Consensus. The ground truth was determined by the interpretion of NCCT images by two US board-certified neuroradiologists, with a third neuroradiologist resolving disagreements.
8. Sample Size for the Training Set
- Training Set Sample Size: Not explicitly stated in the provided text. The document mentions, "Dataset obtained from the RSNA (Radiological Society of North America) Brain CT Hemorrhage Challenge 2019 was used for training and development of the model," but does not specify the exact number of images from this dataset used for training.
9. How the Ground Truth for the Training Set was Established
- Ground Truth Establishment for Training Set: The document states that the "Dataset obtained from the RSNA (Radiological Society of North America) Brain CT Hemorrhage Challenge 2019 was used for training and development of the model." While the specific method of ground truth establishment for that particular dataset isn't detailed here, it implies relying on the ground truth provided with the RSNA challenge dataset, which typically involves expert human annotation.
§ 892.2080 Radiological computer aided triage and notification software.
(a)
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.(b)
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.