(133 days)
Rapid is an image processing software package to be used by trained professionals, including but not limited to physicians and medical technicians. The software runs on a standard off-the-shelf computer or as VMware, and can be used to perform image viewing, processing and analysis of images. Data and images are acquired through DICOM compliant imaging devices.
Rapid provides both viewing and analysis capabilities for functional and dynamic imaging datasets acquired with CT, CT Perfusion (CTP), CT Angiography (CTA), and MRI including a Diffusion Weighted MRI (DWI) Module and a Dynamic Analysis Module (dynamic contrast-enhanced imaging data for MRI and CT).
The CT analysis includes NCCT maps showing areas of hypodense and hyperdense tissue.
The DWI Module is used to visualize local water diffusion properties from the analysis of diffusion - weighted MRI data.
The Dynamic Analysis Module is used for visualization and analysis of dynamic imaging data, showing properties of changes in contrast over time. This functionality includes calculation of parameters related to tissue flow (perfusion) and tissue blood volume.
Rapid CT-Perfusion and Rapid MR-Perfusion can be used by physicians to aid in the selection of acute patients (with known ocuusion of the intracranial internal carotid artery or proximal middle cerebral artery)
Instructions for the use of contrast agents for this in Appendix A of the User's Manual. Additional information for safe and effective drug use is available in the product-specific iodinated CT and gadolinium-based MR contrast drug labeling.
In addition to the Rapid imaging criteria, patients must requirements for thrombectomy, as assessed by the physician, and have none of the following contraindications or exclusions:
- Bolus Quality: absent or inadequate bolus.
- Patient Motion: excessive motion leading to artifacts that make the scan technically inadequate .
- . Presence of hemorrhage
Rapid is a software package that provides for the visualization and study of changes in tissue using digital images captured by diagnostic imaging systems including CT (Computed Tomography) and MRI (Magnetic Image Resonance), as an aid to physician diagnosis. Rapid can be installed on a customer's Server or it can be accessed online as a virtual system. It provides viewing, quantification, analysis and reporting capabilities.
Rapid works with the following types of (DICOM compliant) medical image data:
- CT (Computed Tomography) ●
- MRI(Magnetic Image Resonance)
Rapid acquires (DICOM compliant) medical image data from the following sources:
- . DICOM file
- DICOM CD-R
- Network using DICOM protocol ●
Rapid provides tools for performing the following types of analysis:
- selection of acute stroke patients for endovascular thrombectomy ●
- volumetry of thresholded maps
- time intensity plots for dynamic time courses ●
- measurement of mismatch between labeled volumes on co-registered image ● volumes
- large vessel density
Rapid is a Software as a Medical Device (SaMD) consisting of one or more Rapid Servers (dedicated or virtual). The Rapid Server is an image processing engine that connects to a hospital LAN, or inside the Hospital Firewall. It can be a dedicated Rapid Server or a VM Rapid appliance, which is a virtualized Rapid Server that runs on a dedicated server.
Rapid is designed to streamline medical image processing tasks that are time consuming and fatiguing in routine patient workup. Once Rapid is installed it operates with minimal user interaction. Once the CT (NCCT. CT, CTA) or MR (MR, MRA) data are acquired, the CT or MRI console operator selects Rapid as the target for the DICOM images, and then the operator selects which study/series data to be sent to Rapid. Based on the type of incoming DICOM data, Rapid will identify the data set scanning modality and determine the suitable processing module. The Rapid platform is a central control unit which coordinates the execution image processing modules which support various analysis methods used in clinical practice today:
- Rapid CTP/MRP, DWI, Dynamic Analysis (Original: K121447, Updated with K172477; ● and K182130):
- Rapid CTA (K172477); ●
- Rapid ASPECTS(K190395); ●
- Rapid ICH (K193087);
- . Rapid LVO (K200941);
The iSchemaView Server is a dedicated server that provides a central repository for Rapid data. All iSchemaView Server data is stored on encrypted hard disks. It also provides a user interface for accessing Rapid data. It connects to a firewalled Data Center Network and has its own firewall for additional cyber/data security. The iSchemaView Server connects to one or more Rapid Servers via WAN. Available types of connection include VPN (Virtual Private Network - RFC2401 and RFC4301 Standards) Tunnel and SSH (Secure Shell).
Here's a breakdown of the acceptance criteria and study details for the Rapid device, specifically focusing on the NCCT Motion Artifact AI/ML Module performance, as described in the provided 510(k) summary:
Acceptance Criteria and Reported Device Performance (NCCT Motion Artifact AI/ML Module)
| Metric | Acceptance Criteria (Optimal Performance from training validation) | Reported Device Performance (Final Independent Validation) |
|---|---|---|
| AUC | 0.95 | 0.96 (0.94, 0.97) |
| Sensitivity | 0.95 | 0.91 (0.83, 0.95) |
| Specificity | 0.96 | 0.86 (0.83, 0/89) |
| Primary Endpoint | N/A (implied by meeting sensitivity/specificity targets for "weak artifact = 0") | Passed (weak artifact = 0) |
Study Details
-
Sample sizes used for the test set and the data provenance:
- Test Set Sample Size: N=619 axial image slices.
- Data Provenance: The text does not explicitly state the country of origin for the test set data. It mentions that samples were obtained from "Siemens, GE, Toshiba, Philips, and Neurologica" for training, and for the independent validation, "The samples were primarily from Siemens with GE mixed." This suggests a multi-vendor, and likely multi-site, collection. The study appears to be retrospective as it uses existing medical images for evaluation.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- Number of Experts: 3
- Qualifications of Experts: Described as "experienced truthers." Specific qualifications (e.g., years of experience, subspecialty) are not provided.
-
Adjudication method for the test set:
- The document states "ground truth established by 3 experienced truthers." While it doesn't explicitly mention a 2+1 or 3+1 method, the implication of "established by" multiple experts suggests a consensus-based approach was used to determine the ground truth from these three experts. It does not state "none."
-
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:
- No, a multi-reader, multi-case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance was not conducted or reported in this summary for the NCCT Motion Artifact AI/ML Module. The performance evaluation is for the standalone algorithm.
-
If a standalone (i.e. algorithm only without human-in-the loop performance) was done:
- Yes, a standalone algorithm performance study was done for the NCCT Motion Artifact AI/ML Module. The reported metrics (AUC, Sensitivity, Specificity) are for the algorithm's performance in detecting motion artifacts.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
- The ground truth for the test set was established by expert consensus from 3 experienced truthers.
-
The sample size for the training set:
- Training Set: 23,066 axial image slices (Positive: 1,021, Negative: 12,877).
- Training Validation Set: 5,906 axial image slices (Positive: 422, Negative: 5,484).
-
How the ground truth for the training set was established:
- The document does not explicitly detail how the ground truth for the training data was established. However, given the context of medical image analysis and the subsequent use of "experienced truthers" for independent validation, it's highly probable that human expert review and labeling were also used to establish the ground truth for the training and training validation sets.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).