Search Results

The G-EYE® colonoscope is intended to be used with a compatible video processor (including light source), documentation equipment, monitor, endotherapy device such as a biopsy forceps, and other ancillary equipment for endoscopy and endoscopic surgery within the lower gastrointestinal tract including the anus, rectum, sigmoid colon, colon and ileocecal valve. It incorporates a balloon at the colonoscope to maintain central positioning within the lumen and help control the endoscope's view field and/or endoscope positioning.

The G-EYE® System comprises the G-EYE® Colonoscope and the NaviAid™ SPARKC ("SPARKC") inflation pump. The G-EYE® Colonoscope includes a reprocessable, reusable balloon (the G-EYE® balloon) that is attached to a standard, commercially available colonoscope which is remanufactured by Smart Medical to include the balloon at its bending section. The balloon onto the standard colonoscope is considered remanufacturing of the standard colonoscope by Smart Medical, which assumes responsibility for the final device including the G-EYE® balloon and the remanufactured colonoscope. The G-EYE® balloon is inflated by the dedicated SPARKC inflation system, which enables automatic control of the required balloon pressure in one of several user-selected, system-controlled pressure levels such that the balloon conforms to the colonic lumen diameter. The G-EYE® Colonoscope is advanced by the endoscopist with the balloon deflated, which essentially preserves the standard colonoscope's diameter and insertion technique. At the cecum, the balloon is inflated to engage the colonoscope is then withdrawn by the endoscopist using standard withdrawal technique. Withdrawal of the inflated balloon mechanically straightens the colonic folds and anatomic flexures, centers the colonoscope optics within the colonic lumen, and minimizes uncontrolled bowel slippage during withdrawal, therefore maintaining the endoscope's field of view

The provided text is a 510(k) summary for the G-EYE® System, specifically K202469, which is primarily a submission to expand the range of compatible colonoscope manufacturers. The document repeatedly states that the new device (K202469) is "nearly identical" or "substantially similar" to its predicate device (K192588), and that the differences "do not raise different questions of safety or effectiveness."

Therefore, the document does not describe a new study designed to prove the device meets new acceptance criteria for an AI/ML-based device. Instead, it refers to prior performance testing and clinical studies on the predicate device (K192588) to support the substantial equivalence of the new G-EYE® System (K202469). The "Performance Data" section lists several types of tests, but these are primarily engineering and reprocessing validations, not performance against clinical endpoints that would typically require expert ground truth or MRMC studies.

Based on the provided text, it is not possible to answer the detailed questions regarding acceptance criteria, study design for AI/ML performance, ground truth establishment, or sample sizes for AI/ML model training/testing, as the document focuses on the mechanical and reprocessing aspects of a colonoscopy assist device, and its substantial equivalence to a prior cleared version.

Here's a breakdown of what can be extracted and what cannot:

1. A table of acceptance criteria and the reported device performance:

• The document states that "Each of these tests met its acceptance criteria, supporting substantial equivalence to the predicate." However, the specific acceptance criteria (e.g., numerical targets for performance metrics) are not detailed within this 510(k) summary. The performance reported is qualitative: the tests "met its acceptance criteria" and "demonstrated that the software performs as intended."
• The clinical performance is described as "the device functioned as intended, with no serious device-related adverse events. In addition, the studies demonstrated that the device served its intended function, and that the mechanical straightening of the and anatomic flexures provided by the G-EYE® System facilitates visualization of the colonic mucosa surface area." No specific metrics like sensitivity, specificity, or ADR improvement are provided.

2. Sample sizes used for the test set and the data provenance:

• The document states: "clinical data is available from four clinical studies that further support the safety and performance of the device for its intended use. In total, the studies included approximately 4000 subjects."
• Sample size for test set: The "4000 subjects" refers to participants in prior clinical studies that support the predicate device, not a specific test set for an AI/ML model. It's unclear how many of these were used as a "test set" in the AI/ML sense, or if a separate, dedicated "test set" was used for the current submission's minor changes.
• Data provenance: Not specified (e.g., country of origin, retrospective/prospective). These studies were likely conducted for the predicate device's clearance.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• Not specified. The clinical studies mentioned focused on the device's mechanical function and safety, not on AI/ML performance requiring expert ground truth in the context of diagnostic accuracy.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set:

• Not applicable/Not specified, as this is not an AI/ML diagnostic device being evaluated for standalone or assisted performance against expert ground truth.

5. 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:

• Not mentioned. This device assists withdrawal by straightening folds, not primarily by providing AI-assisted detection or diagnosis. Therefore, an MRMC study comparing human readers with and without AI assistance is not described or relevant for this specific device's function as presented. The product description emphasizes mechanical straightening and improved visualization, not AI-driven lesion detection.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:

• Not applicable. This device is a mechanical aid to colonoscopy, not a standalone diagnostic algorithm.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

• For the clinical studies, "ground truth" would likely be based on standard colonoscopy findings (e.g., visual identification of polyps, histopathology of removed tissues) and patient outcomes. The text primarily focuses on the device's ability to "facilitate visualization" and its safety, rather than providing diagnostic outputs that require "ground truth" in the AI/ML sense.

8. The sample size for the training set:

• Not specified. There is no indication that this device uses machine learning that requires a "training set" in the conventional sense. The "software verification and validation" mentioned likely pertains to the control software for the balloon inflation, not an AI/ML detection algorithm.

9. How the ground truth for the training set was established:

• Not applicable, as no training set for an AI/ML model is indicated.

In summary, the provided 510(k) summary is for a mechanical colonoscopy assist device seeking clearance for expanded compatibility, not an AI/ML diagnostic or detection device requiring the specific types of performance studies listed in the prompt. The "Performance Data" section discusses engineering verification, validation (reprocessing, biocompatibility, software, electrical safety), and references prior clinical studies on the predicate device for overall safety and performance (i.e., its ability to mechanically aid visualization), rather than AI/ML specific metrics.

Ask a specific question about this device