LogoFDA 510(k) Search
Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K192531
    Device Name
    QyScore Software
    Manufacturer
    Qynapse
    Date Cleared
    2019-12-13

    (88 days)

    Product Code
    LLZ
    Regulation Number
    892.2050
    Type
    Traditional
    Panel
    Radiology
    Reference & Predicate Devices
    K170981
    Why did this record match?
    Applicant Name (Manufacturer) :

    Qynapse

    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    QyScore is intended for automatic labeling, visualization and volumetric quantification of segmentable brain structures and lesions from a set of MR images. Volumetric data may be compared to reference percentile data. QyScore is not intended for use in clinical scenarios that require evaluation of the white matter hyperintensities.

    Device Description

    QyScore automatically provides segmentations and measures of brain structures and lesions from a set of MR images for patients between the ages of 20 and 90.

    The software retrieves DICOM MRI data (3DT1 and T2FLAIR series) from a DICOM server and sends it to an analysis server for automatic segmentation of grey matter, white matter, hippocampus, amygdala and white matter hyperintensities. The outputs of the software include an electronic report and color overlays of the segmentation on the input images.

    The results are displayed in a dedicated graphical user interface, allowing the user to:

    • . Browse the segmentations and the measures,
    • Compare the results of segmented brain structures to a reference healthy population, ●
    • Read and edit a PDF report.

    QyScore integrates with leading RIS/PACS systems and can be operated with any MRI scan from 1.5T and 3T scanners for T1 MRI processing, and 3T scanners for T2FLAIR MRI processing.

    AI/ML Overview

    Here's an analysis 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:

    Assessment CategoryAcceptance Criteria (Implicit)Reported Device Performance
    Segmentation Accuracy- Dice coefficients > 85% for whole brain regions (3DT1 MRI scans)- Dice coefficients exceed 85% on whole brain regions (3DT1 MRI scans)
    - Dice coefficients 75%-85% for subcortical brain structures (3DT1 MRI scans)- Dice coefficients are in the range of 75%-85% for subcortical brain structures (3DT1 MRI scans)
    Lesion Segmentation Accuracy- (Implicitly, a low mean absolute volume difference for lesions)- Mean absolute volume difference for lesions: 3.34mL

    2. Sample Size Used for the Test Set and Data Provenance:

    The document mentions that the lesion segmentation was assessed "on a broad validation population, with lesion loads ranging from 0.09mL to 87.65mL." However, the exact sample size (number of cases/patients) for the test set is NOT explicitly stated.

    The data provenance is also NOT explicitly stated (e.g., country of origin, retrospective or prospective).

    3. Number of Experts Used to Establish Ground Truth and Qualifications:

    The document mentions "expert manual segmentations" were used as ground truth. However, the number of experts and their specific qualifications (e.g., radiologist with X years of experience) are NOT explicitly stated.

    4. Adjudication Method for the Test Set:

    The document states that "QyScore segmentations were compared to expert manual segmentations." This implies a direct comparison, but the adjudication method (e.g., 2+1, 3+1, none for discordances) is NOT explicitly stated. It appears to be a direct comparison against a single expert's "manual segmentation" or an agreed-upon expert ground truth.

    5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

    A Multi-Reader Multi-Case (MRMC) comparative effectiveness study was NOT done according to the provided text. The document focuses on the standalone performance of the QyScore software against expert manual segmentations. There's no mention of human readers using or not using the AI and their performance improvement.

    6. Standalone (Algorithm Only) Performance:

    Yes, a standalone (algorithm only without human-in-the-loop performance) study was done. The "accuracy evaluation" explicitly compared "QyScore segmentations to expert manual segmentations" to establish the device's inherent performance.

    7. Type of Ground Truth Used:

    The ground truth used was expert manual segmentations.

    8. Sample Size for the Training Set:

    The sample size for the training set is NOT explicitly stated in the provided text. The document only mentions "a dynamic probabilistic neuroanatomical atlas" used in the segmentation process, which is likely built from a training set, but its size is not detailed.

    9. How the Ground Truth for the Training Set Was Established:

    The document states that the predicate device uses "a dynamic probabilistic neuroanatomical atlas, with age and gender specificity, based on the MR image intensity." The subject device (QyScore) also uses "a dynamic probabilistic neuroanatomical atlas based on the MR image intensity." This implies that the atlas, which serves as a form of ground truth for training, was established through probabilistic modeling of neuroanatomical structures based on a dataset of MR images and their intensities, likely involving expert anatomical knowledge or segmentations used to build the atlas. However, the specific methodology and the process of establishing the ground truth for this training atlas are NOT explicitly detailed.

    Ask a Question

    Ask a specific question about this device

    Page 1 of 1