Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K984443
    Device Name
    ORBSCAN
    Date Cleared
    1999-03-05

    (81 days)

    Product Code
    Regulation Number
    886.1850
    Reference & Predicate Devices
    Predicate For
    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Orbscan II Keratometer is intended to scan, map and display the geometry of the anterior segment of the eye.

    Device Description

    The Orbscan II is a non-invasive, diagnostic system that measures and displays the anterior segment geometry of the eye. The system consists of an optical head, power supply and CPU. The optical head projects light through moving slit shaped apertures, The slit images are projected onto the eye at equi-distant angles from a central optical axis. A video camera placed on the optical axis records the images of the slits as they pass over the eye. The location of each slit at each position is determined during the system calibration.

    During an examination, the slits are stepped across the cornea and an image of the cornea at each slit location is recorded, The CPU processes the images by using edge detection algorithms to determine the location of all of the edges the light strikes during the exam. Through direct triangulation, the elevation and curvature for the various surfaces is determined. Corneal thickness, anterior chamber depth and clevation of the iris and lens can then be determined by subtracting the surfaces.

    This anterior segment information is displayed graphically to a user in the form of colored maps. These maps display variations in height and thickness numerically as well as with gradations in color. The graded colors allow for easier visualization of true measurements and how they relate to elevation and thickness of the eye geometry. The maps can be displayed individually or in several combinations. They can also be customized by the user for their specific examination needs. The Orbscan operates on a Windows™ based user interface and operating system that allows for ease of operation and system control. The CPU used with the Orbscan is a standard, high-speed PC type computer. Other peripherals may include a display monitor, a mouse pointing device and a printer for hard copies of the maps, should the user desire them.

    To enhance repeatability, an optional placido pattern can be attached to the Orbscan. The placido pattern provides slope data for the anterior surface of the eye. The slope data obtained from using the placido pattern is combined with the elevation data as measured with triangulation. Tests on both test objects and human subjects have shown that the reproducibility has been enhanced from +/- . 005mm without the placido to +/-.003 mm with the placido pattern.

    The device is non-invasive and only contacts the patient on his/her chin and forehead. The other concerns for safety are the light output and electrical safety. The light output is of an evesafe intensity and wavelength. Electrical safety is provided by packaging which is compliant to accepted safety standards, and mcdically safe power sources which isolate the patient from hazardous voltages and current.

    The device is proven effective in incasuring the anterior geometry of the eve through internal company and independent clinical studies.

    AI/ML Overview

    Here's an analysis of the Orbscan II Keratometer's acceptance criteria and the study reported to show it meets those criteria, based on the provided document:


    Orbscan IITM Keratometer Performance Study Analysis (K984443)

    1. Acceptance Criteria and Reported Device Performance

    The primary performance metric highlighted and improved upon for the Orbscan II is its reproducibility.

    Acceptance Criteria / Performance MetricPredicate Device Performance (Orbscan)Orbscan II Reported PerformanceNotes
    Reproducibility+/- 0.025 mm+/- 0.003 mmThe improved reproducibility of +/- 0.003 mm is achieved with the optional placido pattern attached. Without the placido, the reproducibility is stated as +/- 0.005mm. The predicate device's figure is likely without a placido equivalent.
    Field of view10 x 14 mmSameNo change noted.
    Axis range0 to 360 degreesSameNo change noted.
    Dioptric range9 to 99 DioptersSameNo change noted.
    Resolution0.10 DioptersSameNo change noted.
    Processing Time~3-5 min (anterior corneal surface)< 1.5 min (all information)Not explicitly an "acceptance criterion" but a significant performance improvement due to increased CPU capability.

    The acceptance criterion, though not explicitly stated as a numerical 'pass/fail' threshold, is implicitly defined by demonstrating a significant improvement in reproducibility over the predicate device, specifically achieving +/- 0.003 mm with the placido pattern.

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

    The document states: "Tests on both test objects and human subjects have shown that the reproducibility has been enhanced from +/- . 005mm without the placido to +/-.003 mm with the placido pattern."

    • Sample Size (Test Set): The exact number of "test objects" and "human subjects" used is not specified in the provided summary.
    • Data Provenance: The document does not specify the country of origin. It indicates that the studies were "internal company and independent clinical studies," implying a mix of sources. It is likely retrospective or prospective depending on the nature of the "internal company" versus "independent clinical studies," but this is not explicitly stated for the reproducibility data itself.

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

    The document does not provide information on the number or qualifications of experts used to establish ground truth for the test set.

    4. Adjudication Method for the Test Set

    The document does not specify any adjudication method (e.g., 2+1, 3+1, none) for the test set. Given the nature of reproducibility measurements (instrument measuring the same object multiple times), traditional expert adjudication for clinical agreement might not be applicable in the same way as for diagnostic interpretation.

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

    There is no mention of a Multi-Reader Multi-Case (MRMC) comparative effectiveness study being performed. The device is for measurement and display of geometry, not for interpretation by multiple human readers in a diagnostic context that would typically warrant an MRMC study related to interpretation improvement.

    6. Standalone (Algorithm Only) Performance

    The primary performance metric (reproducibility) implicitly describes the standalone performance of the algorithm and the device system. The device captures data and processes it using "edge detection algorithms" and "direct triangulation" to determine measurements. The reproducibility figures reflect the consistency of these automated measurements. The improved processing time (less than 1.5 min) also relates to the standalone performance of the algorithm and integrated system.

    7. Type of Ground Truth Used

    The ground truth for the reproducibility study would be the physical dimensions and geometry of the "test objects" or the "true" anterior segment geometry of the human subjects. This would typically be established through highly accurate, independent measurement methods, though these specific methods are not detailed. For human subjects, repeated measurements over a short period with minimal physiological change serve as the basis for assessing instrument reproducibility.

    8. Sample Size for the Training Set

    The document does not specify a sample size for any training set. It mentions "edge detection algorithms" and "direct triangulation," which are often based on well-established mathematical principles and might not involve a machine learning "training set" in the modern sense. If any training was involved for internal algorithm development, it is not disclosed.

    9. How Ground Truth for the Training Set Was Established

    As no training set is explicitly mentioned, the method for establishing its ground truth is also not described. For rule-based algorithms, ground truth might be implicitly built into the mathematical models rather than established through labeled datasets.

    Ask a Question

    Ask a specific question about this device

    Page 1 of 1