(56 days)
EFAI Chestsuite XR Pleural Effusion Assessment System is a software workflow tool designed to aid the clinical assessment of adult (18 years of age or older) Chest X-Ray cases with features suggestive of pleural efflusion in the medical care environment. EFAI Chestsuite XR Pleural Effusion Assessment System analyzes cases using an artificial intelligence algorithm to identify suspected findings on chest x-ray images taken in PA position. It makes case-level output available to a PACS/workstation for worklist prioritization or triage. EFAI Chestsuite XR Pleural Effusion Assessment System is not intended to direct attention to specific portions or anomalies of an image. Its results are not intended to be used on a stand-alone basis for clinical decision-making nor is it intended to rule out pleural effusion or otherwise preclude clinical assessment of X-Ray cases.
EFAI ChestSuite XR Pleural Effusion Assessment System, is a radiological computer-assisted triage and notification software system. The software uses deep learning techniques to automatically analyze PA chest x-rays and sends notification messages to the picture archiving and communication system (PACS)/workstation to allow suspicious findings of pleural effusion to be identified.
The device is intended to provide a passive notification through the PACS/workstation to the radiologists indicating the existence of a case that may potentially benefit from the prioritization. It does not mark, highlight, or direct users' attention to a specific location on the original chest X-ray. The device aims to aid in prioritization and triage of radiological medical images only.
The deployment environment is recommended to be in a local network with an existing hospitalgrade IT system in place. EFAI Chestsuite XR Pleural Effusion Assessment System should be installed on a specialized server supporting deep learning processing. The configurations are only being operated by the manufacturer:
- Local network setting of input and output destinations; ●
EFAI Chestsuite XR Pleural Effusion Assessment System is a software-only device which operates in four stages - data transfer, data preprocessing. AI inference and data post processing. The workflow of the device begins with applying a number of filtering rules based on image characteristics and DICOM tags to ensure only eligible images are analyzed by the algorithm. The image preprocessing unit ensures that all the input data are normalized (image size, contrast) such that it is ready for the algorithm to conduct the analysis. The AI inference generates an assessment which is then post-processed into a JSON message and transferred to an API server. The software is packaged as a docker container such that it can be installed and deployed to both a physical or virtual machine.
Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
| Metric | Acceptance Criteria (Lower Bound) | Reported Device Performance (95% CI) |
|---|---|---|
| AUC | > 0.95 | 0.9712 (0.9538-0.9885) |
| Sensitivity | > 0.80 | 0.9510 (0.9195-0.9706) |
| Specificity | > 0.80 | 0.9745 (0.9505-0.9870) |
The reported device performance for all metrics (AUC, Sensitivity, Specificity) exceeded their respective acceptance criteria.
2. Sample size used for the test set and the data provenance
- Sample Size: 600 anonymized Chest X-ray images (286 positive for pleural effusion, 314 negative).
- Data Provenance: Retrospective cohort collected from multiple institutions in the US and OUS (Outside US).
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
- Number of Experts: Three.
- Qualifications of Experts: US board-certified radiologists. The specific number of years of experience is not mentioned.
4. Adjudication method for the test set
- Adjudication Method: Majority agreement was used as the reference standard (ground truth). This implies a 3-reader consensus where at least 2 out of 3 had to agree.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
- MRMC Study: No, a multi-reader multi-case (MRMC) comparative effectiveness study with human-in-the-loop performance was not explicitly done or reported in this document. The study focused on the standalone performance of the AI algorithm. Therefore, no effect size of human readers improving with AI vs. without AI assistance is provided.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
- Standalone Performance: Yes, a standalone performance test was performed to compare the pleural effusion classification performance and processing time of the EFAI Chestsuite XR Pleural Effusion Assessment System against the predicate device, HealthCXR.
7. The type of ground truth used
- Type of Ground Truth: Expert consensus (majority agreement of three US board-certified radiologists).
8. The sample size for the training set
- The document mentions an "internal validation test" with 1454 images collected retrospectively between 2006-2018 from Taiwan, where "Ground-truthing (classified into positive and negative of pleural effusion) was done by three board-certified radiologists." This sounds like an internal validation set rather than the training set. The true size of the training set is not explicitly stated in the provided text.
9. How the ground truth for the training set was established
- As the training set size is not explicitly stated, the method for establishing its ground truth is also not explicitly detailed. However, for the internal validation set mentioned (1454 images), the ground truth was established by three board-certified radiologists. It's highly probable that a similar method (expert review) was used for the training data as well, given the nature of the task.
§ 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.