(58 days)
The Vitek GNI+ Card is intended to be used in conjunction with the Vitek System for the automated identification of microorganism of the family Enterobacteriaceae. In addition, a select group of glucose non-fermenting gram-negative bacteria, and members of the family Vibrionaceae can be identified. In addition to the organisms currently claimed with the GNI Card, the GNI+ Card has added the following 13 new organism claims:
- Aeromanas veronii biovar veronii Budvicia aquatica Burkholdenia mallei CDC Group EO-2 Chromobacterium violaceum Edwardsiella hoshinae Klebsiella ornithinolytica Moellerella wisconsensis Rahnella aquatilis Shigella boydii/flexneri Sphingobacterium spiritivorum Sphingobacterium thalpophilum Yokenella regensburgei (Koserella trabulsii)
The GNI+ Card consists of a plastic card with thirty wells that contain 28 dehydrated biochemical broths, one negative control broth and one growth control broth. The broths are re-hydrated with a saline suspension of a pure culture isolated from a patient specimen. The GNI+ Card performs a variety of conventional and non-conventional biochemical tests that are based on established biochemical methods. Organism identification usually requires between 2-12 hours of incubation on the automated Vitek System.
Here's an analysis of the provided 510(k) summary, structured to address your specific questions:
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria are not explicitly stated as numerical targets in this summary. However, based on the presented data, we can infer the implicit performance expectations.
| Acceptance Criteria (Inferred) | Reported Device Performance |
|---|---|
| High rate of correct identification (developmental phase) | 96.9% correctly identified in non-clinical development |
| High rate of correct identification (challenge organisms) | 94.7% correctly identified in clinical challenge study |
| High overall correlation with existing GNI Card | 97.4% overall correlation in clinical trials |
| No adverse discrepancies | "No adverse discrepancies were noted" |
| Software performance meets specifications | "tested extensively in-house" and "at clinical trial sites" to a "detailed protocol" and "designed, validated and tested according to established written procedures" |
2. Sample Size Used for the Test Set and Data Provenance
- Test Set (Clinical Trials - Challenge Organisms): 106 challenge organisms. These were "different isolates than those used to develop the database." The provenance is internal to bioMérieux Vitek.
- Test Set (Clinical Trials - Routine Clinical Isolates): "A minimum of 130 routine clinical isolates." These were "from fresh patient cultures as they were randomly submitted to the laboratory." The country of origin is not specified but implied to be the locations of the three clinical sites. The data is prospective for these isolates.
- Test Set (Clinical Trials - Investigator's Stock Collection): "Thirty to fifty isolates from the investigator's stock collection." The provenance is internal to the clinical sites. This would likely be retrospective, drawing on existing collections.
- Test Set (Non-Clinical Development Trials): 2974 strains were used to develop the GNI+ database. These were "well characterized, having been previously identified." The provenance is internal to bioMérieux Vitek. This data is retrospective.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
The document does not specify the number of experts or their qualifications for establishing ground truth in the clinical trials. It relies on established methods and their outcomes.
4. Adjudication Method for the Test Set
The adjudication method described is a hierarchical approach:
- Initial Comparison: GNI+ card results were directly compared to the current Vitek GNI Card.
- First Level Discrepancy Resolution: Discrepancies between GNI+ and GNI were resolved using:
- API 20E (for fermenters)
- API NFT (for non-fermenters)
- Second Level Discrepancy Resolution: Further discrepancies, or if GNI+ identified a species not claimed by the GNI Card or API methods, were confirmed with "conventional biochemical testing."
This method is essentially a form of expert-guided consensus, where established, validated methods (API strips, conventional biochemicals) serve as the "experts" for adjudication.
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
No, an MRMC comparative effectiveness study involving human readers or AI assistance in that context was not performed. This device is an automated identification system, not an AI-assisted diagnostic tool that aids human interpretation. The "AI" in this context refers to the automated analysis and identification algorithm, not a system designed to improve human reader performance.
6. If a Standalone (i.e. algorithm only without human-in-the-loop performance) Was Done
Yes, a standalone performance evaluation of the algorithm was conducted. The entire premise of the GNI+ Card and Vitek System is automated identification. Both the "Non Clinical Development Trials" (96.9% correct identification) and the "Clinical Trials" (94.7% correct on challenge set, 97.4% overall correlation) represent standalone performance of the device without human interpretation of the raw biochemical reactions. Humans prepare the sample, but the identification itself is automated.
7. The Type of Ground Truth Used (expert consensus, pathology, outcomes data, etc.)
The ground truth for the clinical trials was established through:
- Reference Devices: The current Vitek GNI Card.
- Established Commercial Kits: API 20E and API NFT.
- Conventional Biochemical Testing: Standard laboratory methods for bacterial identification.
This represents a form of expert consensus based on established laboratory methods and reference technologies, rather than direct human expert consensus alone or pathology/outcomes data.
8. The Sample Size for the Training Set
The "Non Clinical Development Trials" section states: "A panel of 2974 strains were used to develop the GNI + database." This panel served as the training set for the algorithm that analyzes the biochemical reactions.
9. How the Ground Truth for the Training Set Was Established
The ground truth for the 2974 strains used in the training set (to develop the GNI+ database) was established by their prior identification using:
- The GNI Card (existing device)
- API strips (commercial reference method)
- Conventional biochemicals (standard laboratory methods)
These organisms were described as "well characterized," implying their identities were confidently established through these recognized methods prior to their use in database development. This also constitutes expert consensus based on established laboratory methods and reference technologies.
§ 866.2660 Microorganism differentiation and identification device.
(a)
Identification. A microorganism differentiation and identification device is a device intended for medical purposes that consists of one or more components, such as differential culture media, biochemical reagents, and paper discs or paper strips impregnated with test reagents, that are usually contained in individual compartments and used to differentiate and identify selected microorganisms. The device aids in the diagnosis of disease.(b)
Classification. Class I (general controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to § 866.9.