(47 days)
Segment is a software that analyzes DICOM-compliant cardiovascular images acquired from magnetic resonance (MR) scanners. Segment specifically analyzes the function of the heart and its major vessels using multi-slice, multi-frame and velocity encoded MR images. It provides clinically relevant and reproducible data for supporting the evaluation of the function of the chambers of the heart such as left and right ventricular volumes, ejection fractions, stroke volumes, peak ejection and filling rates, myocardial mass, regional wall thickness, fractional thickening and wall motion. It also provides quantitative data on blood flow and velocity in the arterial vessels and at the heart valves. Segment is tested on MR images acquired from both 1.5 T and 3 T MR scanners. The data produced by Segment is intended to be used to support qualified cardiologist, radiologist or other licensed professional healthcare practitioners for clinical decision making. It is a support tool that provides relevant clinical data as a resource to the clinician and is not intended to be a source of medical advice or to determine or recommend a course of action or treatment for a patient.
Segment is a software for analysis of cardiovascular MR images. Segment provides clinical quantitative data by analyzing multi-slice, multi-phase DICOM compatible cardiovascular MR images. Functional and blood flow analysis is performed using 2D, 3D and 4D data sets using standard algorithms and user input. MR images may be imported from various sources including images stored on portable media, network storage devices, PACS, and other vendor systems and supports cardiovascular MR images from all of the major MRI scanner vendors. Segment can be used for quantitative and qualitative analysis of cardiovascular MR images.
The provided text is a 510(k) Summary for the Medviso AB Segment software, dating from 2009. It focuses on demonstrating substantial equivalence to predicate devices (MRI-MASS and MRI-FLOW) rather than providing detailed acceptance criteria and a specific study proving the device meets those criteria with statistical significance as typically found in more recent submissions for AI/ML devices.
Here's an analysis based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance:
The document does not explicitly state quantitative acceptance criteria or a table of reported device performance in the way a modern AI/ML device submission would. Instead, the "performance" is implicitly tied to demonstrating substantial equivalence to the predicate devices and general claims of clinical utility and reproducibility.
| Criterion | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Functional Equivalence | Substantially equivalent to predicate devices (MRI-MASS and MRI-FLOW) in analyzing multi-slice, multi-frame, and phase-encoded DICOM-compliant MR images, utilizing standard algorithms and user inputs for delineation of myocardial and arterial vascular walls. | "Segment has substantially equivalent features and specifications to the predicate device." "Segment performs in accordance with its intended use as well as the Medis MRI-MASS and MRI-FLOW cardiovascular MRI image analysis products currently on the market." |
| Clinical Utility (Qualitative) | Provides clinically relevant and reproducible data for supporting evaluation of heart and major vessel function. | "It provides clinically relevant and reproducible data for supporting the evaluation of the function of the chambers of the heart..." "The image analysis provided by Segment makes the images more clinically useful for the physician in making his diagnosis." |
| Clinical Utility (Quantitative) | Provides quantitative data for metrics such as LV/RV volumes, EF, stroke volumes, peak ejection/filling rates, myocardial mass, wall thickness, fractional thickening, wall motion, blood flow, and velocity in vessels/valves. | "It provides clinically relevant and reproducible data for supporting the evaluation of the function of the chambers of the heart such as left and right ventricular volumes, ejection fractions, stroke volumes, peak ejection and filing rates, myocardial mass, reqional wall thickness, fractional thickening and wall motion. It also provides quantitative data on blood flow and velocity in the arterial vessels and at the heart valves." |
| Image Compatibility | Compatible with DICOM-compliant cardiovascular MR images from various sources (portable media, network, PACS, other vendor systems) and tested on 1.5 T and 3 T MR scanners. | "MR images may be imported from various sources including images stored on portable media, network storage devices, PACS, and other vendor systems and supports cardiovascular MR images from all of the major MRI scanner vendors." "Segment is tested on MR images acquired from both 1.5 T and 3 T MR scanners." |
| Safety and Effectiveness (Hazard Mitigation) | All identified hazards mitigated to minor levels of concern. | Hazards identified in Hazard Analysis (section 16.3) and controlled by design controls, protective measures, and warnings to users, mitigating them to "minor levels of concern." |
| Interoperability | Runs on standard PC hardware and Microsoft Windows XP or Vista. | "Segment runs on standard PC hardware and Microsoft Windows XP or Microsoft Windows Vista operating system software." |
2. Sample Size Used for the Test Set and Data Provenance:
The document does not specify a sample size for any formal "test set" in the context of validating performance against ground truth. The submission predates the common expectations for rigorous AI/ML performance validation studies. It states that "Extensive testing of the software package is performed by programmers, by non-programmers, quality assurance staff and by potential customers prior to commercial release." This refers to internal software quality assurance and user acceptance testing, not a clinical performance validation study with a defined test set and ground truth.
There is no mention of data provenance (e.g., country of origin or retrospective/prospective nature) for any data used in testing.
3. Number of Experts Used to Establish Ground Truth and Qualifications:
The document does not mention using experts to establish ground truth for a test set. This type of submission relies on the established clinical utility of MRI images and the predicate devices' method of analysis, with the software acting as a tool to aid qualified professionals. The output is "intended to be used to support qualified cardiologist, radiologist or other licensed professional healthcare practitioners."
4. Adjudication Method:
Given that no formal expert-derived ground truth test set is described, there is no mention of an adjudication method (e.g., 2+1, 3+1).
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
No, a MRMC comparative effectiveness study was not done (or at least, not reported in this 510(k) summary). The submission focuses on substantial equivalence based on features and functionality, not on comparing reader performance with and without AI assistance. Therefore, no effect size calculation for human reader improvement with AI assistance is provided.
6. Standalone (Algorithm Only) Performance Study:
No, a standalone performance study was not explicitly described in the context of quantitative performance metrics against a defined ground truth. The document states, "Segment is used for image analysis and quantification... and all of the information is subject to his/her oversight and control." This indicates that the device is intended as a support tool, not a fully autonomous diagnostic algorithm. The "extensive testing" mentioned is internal software testing and potential customer feedback, not a formal standalone clinical performance study.
7. Type of Ground Truth Used:
No specific "ground truth" (e.g., pathology, outcomes data, or expert consensus specifically for a test set) is described as being used to validate the device's performance. The basis for safety and effectiveness is largely on clinical acceptance of MRI imaging itself and substantial equivalence to existing, legally marketed devices. The software "does not in any way alter the images" and provides "clinically relevant numeric computations."
8. Sample Size for the Training Set:
The document does not mention a training set or its sample size. This is consistent with the era and type of device; Segment is an image analysis software based on "standard algorithms and user input," rather than a machine learning model that requires a "training set" in the modern sense.
9. How the Ground Truth for the Training Set Was Established:
As no training set is discussed, no information is provided on how its ground truth would have been established.
§ 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).