LogoFDA 510(k) Search
Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K083423
    Device Name
    COLONCAD API 3.1
    Manufacturer
    MEDICSIGHT PLC
    Date Cleared
    2011-05-17

    (909 days)

    Product Code
    NWE
    Regulation Number
    892.2050
    Type
    Traditional
    Panel
    Radiology (RA)
    Reference & Predicate Devices
    K042674,K042605
    Why did this record match?
    Applicant Name (Manufacturer) :

    MEDICSIGHT PLC

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

    Medicsight ColonCAD™ API is a non-invasive computer aided detection (CAD) image analysis software tool designed to assist radiologists in the detection of colorectal polyps during their review of digital images derived from CT colonography (CTC). This CAD software post-processes the CTC image data obtained from multi-detector computed tomography (MDCT) scanners.

    The device is intended to be used on patients referred for a CT Colonography examination, as an overlay tool to prompt the radiologist to colonic findings that have been identified by the device. The CAD can assist radiologists after they have made an initial review of all the CTC image data, supporting their evaluation ("second read").

    Device Description

    Medicsight ColonCAD API is a medical imaging software tool designed to assist radiologists in the detection of polyps in CT scans of the product is packaged as an Application Programming Interface (API) which allows it to be integrated into existing medical imaging solutions.

    The ColonCAD API assists the radiologist in detecting colorectal polyps using mathematical image processing techniques. The CAD assists the radiologist by highlighting potential polyps in 2D and 3D image views. The results are displayed in the form of "CAD marks" on or near the potential polyps. The radiologist must assess every CT scan image to search for polyps and review the CAD marked images to determine if the indicated findings are polyps.

    Patient management decisions should not be made solely on the results of ColonCAD analysis.

    AI/ML Overview

    The information provided details the Medicsight ColonCAD API device, its intended use, and a summary of studies conducted. However, the document does not explicitly state specific acceptance criteria (e.g., a required sensitivity or specificity value) or the reported device performance in a numerical table. It only mentions that the device's accuracy was "significantly higher" with CAD assistance.

    Therefore, I cannot populate a table of acceptance criteria and reported device performance with specific numerical values based on this document.

    Here's an analysis of the provided information concerning the study:

    1. Table of Acceptance Criteria and Reported Device Performance

    Acceptance CriteriaReported Device Performance
    Not explicitly stated in the provided document.Radiologists' accuracy for detecting colorectal polyps of any size was significantly higher with CAD than in the unassisted read, as measured by the segment-level area under the ROC curve (AUC).

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

    • Sample Size for Test Set: Not explicitly stated in the provided document.
    • Data Provenance: Not explicitly stated in the provided document (e.g., country of origin, retrospective or prospective).

    3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

    • Number of Experts: Not explicitly stated in the provided document.
    • Qualifications of Experts: Not explicitly stated in the provided document.

    4. Adjudication Method for the Test Set

    • Adjudication Method: Not explicitly stated in the provided document.

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

    • Was an MRMC study done? Yes. The document states: "The results of the MRMC study demonstrated that radiologists' accuracy for detecting colorectal polyps of any size was significantly higher with CAD than in the unassisted read, as measured by the segment-level area under the ROC curve (AUC)."
    • Effect Size: The document states that accuracy was "significantly higher" with CAD, and mentions "segment-level area under the ROC curve (AUC)" as the metric. However, it does not provide a numerical effect size (e.g., the specific AUC values for assisted vs. unassisted read, or the magnitude of improvement).

    6. Standalone Performance Study

    • Was a standalone study done? The document focuses on the C-CAD (Computer Aided Detection) device assisting radiologists, implying a human-in-the-loop scenario. While it mentions "internal clinical evaluations" as part of non-clinical studies, it does not detail a standalone algorithm-only performance study with specific metrics. The primary clinical study discussed is an MRMC study evaluating human-in-the-loop performance.

    7. Type of Ground Truth Used

    • Type of Ground Truth: Not explicitly stated in the provided document for the clinical study. It refers to "detecting colorectal polyps," implying a definitive diagnosis, but the method for establishing this truth is not detailed (e.g., pathology, expert consensus).

    8. Sample Size for the Training Set

    • Sample Size for Training Set: Not explicitly stated in the provided document. The document mentions the ColonCAD API uses the "same underlying image processing technology" and "same algorithm" as the predicate ColonCAR 1.2 device (K042674), suggesting the training might have occurred prior to this specific API version's development or relies on pre-trained models.

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

    • How Ground Truth Was Established (Training Set): Not explicitly stated in the provided document. As with the test set, the method for establishing the ground truth for training data (if new training was performed) is not detailed.
    Ask a Question

    Ask a specific question about this device

    Page 1 of 1