(345 days)
For the in-vitro measurement of autoantibodies against B2-glycoprotein 1 (B2GP1) present in human serum. This kit may be used in conjunction with anticardiolipin assays and clinical information to aid the diagnosis of thrombosis in at risk patients having, for example, antiphospholipid syndrome (APS) or systemic lupus erythematosus (SLE).
Not Found
This FDA letter (K991803) is an approval for a diagnostic test kit (Bindazyme® Anti-B2 GP1 IgM EIA Test Kit), not an AI/ML medical device. Therefore, the traditional acceptance criteria and study design elements requested in the prompt, such as those related to AI model performance, expert consensus, and reader studies, are not applicable here.
However, I can extract the available information from the document that is relevant to diagnostic device approval.
Acceptance Criteria and Device Performance (Based on information typically available for EIA test kits):
For an EIA test kit like the Bindazyme® Anti-B2 GP1 IgM EIA Test Kit, acceptance criteria focus on its analytical performance (how well it measures what it's supposed to measure) and clinical performance (how accurately it identifies disease states).
Since this document is only the FDA clearance letter, it does not contain the detailed performance data or specific acceptance criteria met internally by the manufacturer. However, for an in vitro diagnostic device seeking substantial equivalence, key performance characteristics typically evaluated and accepted would include:
| Acceptance Criteria Category | Typical Performance Metric | Reported Device Performance (Not explicitly in document, inferred for K991803) |
|---|---|---|
| Analytical Performance | Sensitivity | Not specified in this document. Would typically measure the lowest concentration of analyte that can be reliably detected. |
| Specificity | Not specified in this document. Would typically assess interference from other substances. | |
| Precision/Reproducibility | Not specified in this document. Would evaluate the consistency of results when the same sample is tested multiple times. | |
| Linearity | Not specified in this document. Would assess if the test gives proportionally accurate results across a range of analyte concentrations. | |
| Inter-Assay Variability | Not specified in this document. Would assess consistency across different test runs. | |
| Cross-Reactivity | Not specified in this document. Would confirm the test does not react with unrelated substances. | |
| Clinical Performance | Clinical Sensitivity | Not specified in this document. For a diagnostic test, this would be the percentage of actual disease cases correctly identified by the test. |
| Clinical Specificity | Not specified in this document. For a diagnostic test, this would be the percentage of actual non-disease cases correctly identified by the test. | |
| Positive Predictive Value | Not specified in this document. Would be the probability that a positive test result indicates disease. | |
| Negative Predictive Value | Not specified in this document. Would be the probability that a negative test result indicates no disease. | |
| Agreement with Predicate | Substantially Equivalent (as stated in the letter). The FDA determined the device's performance was comparable to legally marketed predicate devices, implying similar accuracy. |
Study Information (Based on typical requirements for K99 clearance):
-
Sample size used for the test set and the data provenance: Not explicitly stated in the FDA clearance letter. For in vitro diagnostic devices, comparative studies would have been conducted using patient samples (often both retrospective and prospective) to establish performance relative to a predicate device and/or clinical gold standard. The country of origin for the data is also not specified.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts: Not applicable in the context of an in vitro diagnostic assay like this. The "ground truth" for such an assay is established by clinical diagnosis, often through a combination of other laboratory tests, clinical presentation, and expert physician judgment, but not typically in the same way an AI ground truth is established by a panel of image readers. For the assay itself, the "ground truth" samples would be derived from patients with confirmed clinical diagnoses (e.g., confirmed APS or SLE with known B2GP1 status via other methods).
-
Adjudication method for the test set: Not applicable. Adjudication methods like 2+1 or 3+1 are used for expert consensus on AI ground truth. For an in vitro diagnostic device, sample classification would be based on established clinical diagnostic criteria or comparison to a predicate device.
-
If a multi-reader multi-case (MRMC) comparative effectiveness study was done: No. MRMC studies are specific to imaging interpretation by human readers, often comparing AI-assisted vs. unassisted performance. This is an in vitro diagnostic test kit, not an imaging device or an AI assistant for human interpretation.
-
If a standalone (i.e. algorithm only without human-in-the-loop performance) was done: The device itself is a "standalone" laboratory test in the sense that it produces a result without human interpretation of complex visual patterns in the same way an AI algorithm analyzes images. Its performance is evaluated on its ability to accurately measure the target analyte and correlate with clinical conditions.
-
The type of ground truth used: For a diagnostic assay like this, the "ground truth" for validating its performance would typically come from:
- Clinical Diagnosis: Patients definitively diagnosed with/without conditions like Antiphospholipid Syndrome (APS) or Systemic Lupus Erythematosus (SLE) based on established clinical criteria and other definitive tests.
- Reference Methods/Predicate Devices: Results from well-established, previously cleared tests for anti-B2GP1 IgM or related markers.
- Pathology/Outcomes Data: In some cases, long-term patient outcomes might contribute to affirming diagnostic accuracy, but for a serological test, this is usually indirect.
-
The sample size for the training set: Not applicable for this type of device. "Training set" refers to data used to train an AI/ML algorithm. For an in vitro diagnostic assay, the manufacturer develops and optimizes the assay parameters (e.g., reagents, concentrations, incubation times) through extensive R&D and optimization studies, not through a "training set" in the AI sense.
-
How the ground truth for the training set was established: Not applicable for the reasons mentioned above. Assay optimization and development use samples and experiments to refine the assay's performance characteristics, rather than establishing "ground truth" for a training set.
§ 866.5660 Multiple autoantibodies immunological test system.
(a)
Identification. A multiple autoantibodies immunological test system is a device that consists of the reagents used to measure by immunochemical techniques the autoantibodies (antibodies produced against the body's own tissues) in serum and other body fluids. Measurement of multiple autoantibodies aids in the diagnosis of autoimmune disorders (disease produced when the body's own tissues are injured by autoantibodies).(b)
Classification. Class II (performance standards).