(96 days)
Brainomix 360 Triage Stroke is a radiological computer aided triage and notification software indicated for use in the analysis of non-contrast head CT (NCCT) images to assist hospital networks and trained clinicians in workflow triage by flagging and communicating suspected positive findings of head NCCT images for large vessel occlusion (LVO) of the intracranial ICA and M1 and intracranial hemorrhage (ICH). Specifically, the device is intended to be used for the trage of images acquired from adult patients in the acute setting, within 24 hours of the acute symptoms, or where this is unclear, since last known well (LKW) time. It is not intended to detect isolated subarachnoid hemorrhage and symmetrical bilateral MCA occlusions.
Brainomix 360 Triage Stroke uses an artificial intelligence algorithm to analyze images and highlight cases with detected NCCT LVO or ICH on the Brainomix server on premise or in the cloud in parallel to the ongoing standard of care image interpretation. The user is presented with notifications for cases with suspected LVO or ICH findings via a web user interface or mobile applications include compressed preview images that are meant for informational purposes only and are not intended for diagnostic use beyond notification.
The device does not alter the original mage, and it is not intended to be used as a primary diagnostic device. The results of Brainomix 360 Triage Stroke 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 ultimately responsible for reviewing full images per the standard of care.
Brainomix 360 Triage Stroke is a radiological computer aided triage and notification software package compliant with the DICOM standard and running on an off-the-shelf physical or virtual server.
The Triage Stroke module is a non-contrast CT processing module which operates within the integrated Brainomix 360 Platform to provide triage and notification of suspected large vessel occlusion (LVO) and intracranial hemorrhage (ICH). Brainomix 360 Triage Stroke is a stand-alone software device which uses machine learning algorithms that uses advanced non adaptive imaging artificial intelligence, and large data analytics to automatically identify suspected LVO and ICH on non-contrast CT (NCCT) imaging acquired from adult patients in the acute setting, within 24 hours of the acute symptoms, or where this is unclear, since last known well (LKW) time. The output of the module is a priority notification to clinicians indicating the suspicion of LVO or ICH based on positive findings. Specifically, Brainomix 360 Triage Stroke's ICH analysis is optimized to identify findings of hyperdense volume in the parenchyma typically associated with acute intracranial hemorrhage; and the NCCT LVO suspicion uses the combined analysis of the ASPECTS and hyperdense vessel sign (HDVS) algorithms. It is not intended to detect isolated subarachnoid hemorrhage and symmetrical bilateral MCA occlusions. The Triage Stroke module uses the basic services supplied by the Brainomix 360 Platform including DICOM processing, job management, imaging module execution and imaging output including the notification and compressed image.
Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:
Acceptance Criteria and Device Performance
Device Name: Brainomix 360 Triage Stroke
Indications for Use: Radiological computer aided triage and notification software for analysis of non-contrast head CT (NCCT) images to assist workflow triage by flagging and communicating suspected positive findings of large vessel occlusion (LVO) of the intracranial ICA and M1 and intracranial hemorrhage (ICH).
| Acceptance Criteria Category | Specific Metric | Acceptance Criteria (Target) | Reported Device Performance |
|---|---|---|---|
| ICH Detection | Sensitivity | Exceeded pre-specified goals | 92.5% (95% Cl: 80.97-98.36%) |
| Specificity | Exceeded pre-specified goals | 87.22% (95% Cl: 82.39-91.18%) | |
| NCCT LVO Detection | Sensitivity | Exceeded pre-specified goals | 68.75% (95% Cl: 59.71-76.90%) |
| Specificity | Exceeded pre-specified goals | 89.57% (95% Cl: 82.92-94.36%) | |
| Time-to-Notification | Total Time | Under 3.5 minutes | Minimum: 62 seconds, Maximum: 134 seconds (Met criteria) |
Study Details
1. Sample Size Used for the Test Set and Data Provenance
- Test Set Sample Size: 267 cases (40 ICH positive, 112 LVO positive, 115 negative for ICH or LVO, 3 excluded due to technical inadequacy).
- Data Provenance: Retrospective study. The country of origin is not explicitly stated, but the company is based in the United Kingdom.
2. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications
- Number of Experts: Three (for LVO cases). The number of experts for ICH cases is implied by "previously truthed," likely referring to the K231195 submission, but not specified in this document.
- Qualifications of Experts: Experienced US board-certified neuroradiologists.
3. Adjudication Method for the Test Set
- Adjudication Method: Consensus (for LVO cases). For ICH cases, the method is "as described in the standalone study for our previously cleared device," but not detailed here.
4. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done
- Was an MRMC study done? Yes, a reader study was conducted to compare NCCT LVO sensitivity of the device to that of radiologists.
- Effect Size of Human Readers Improvement with AI vs. Without AI Assistance:
- The study compared the device's standalone sensitivity to the sensitivity of human readers. It did not directly measure how human readers improve with AI assistance (i.e., human-in-the-loop performance with AI vs. without AI).
- Device's standalone sensitivity: 68.75%
- All readers (experts and non-experts) sensitivity: 47.94% (95% Cl: 37.91-57.97%)
- Difference between device's sensitivity and all readers: 20.52% (95% Cl: 8.26-32.78%)
- General radiologists (non-experts) sensitivity: 47.18% (95% Cl: 33.62-60.75%)
- Difference between device and non-expert sensitivity: 21.28% (95% Cl: 5.84-36.72%)
- The study stated that the device passed "expert non-inferiority and non-expert superiority," implying the device performs at least as well as experts and better than non-experts in terms of sensitivity for LVO detection.
5. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
- Was a standalone study done? Yes. The core performance data for ICH and NCCT LVO (sensitivity and specificity) represent the algorithm's standalone performance.
6. The Type of Ground Truth Used
- ICH Cases: NCCT imaging with additional clinical information (as described in a previous submission for a related device).
- LVO Cases: Acute CTA imaging and additional clinical information.
- Method of Ground Truth Establishment: Expert consensus (for LVO cases).
7. The Sample Size for the Training Set
- The document does not specify the sample size for the training set. It mentions the algorithm "uses advanced non adaptive imaging artificial intelligence, and large data analytics," which implies a training phase, but no details on size are provided.
8. How the Ground Truth for the Training Set was Established
- The document does not explicitly state how the ground truth for the training set was established. It only describes the ground truth establishment for the test set.
§ 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.