(73 days)
ImmunoCard STAT!® RSV is a rapid, qualitative, lateral-flow immunoassay for the detection of Respiratory Syncytial Virus (RSV) antigens (fusion protein and internal protein) in human Nasal Wash, Nasopharyngeal Aspirate and Nasal and Nasopharyngeal Swab Samples. It is designed to test specimens from symptomatic neonatal and pediatric patients. It is recommended that all negative test results be confirmed by cell culture.
ImmunoCard STAT! RSV is distributed as a test kit that includes the following reagents: ImmunoCard STAT! RSV Test Device: A chromatography strip housed in a plastic frame and enclosed in a foil pouch with a desiccant. The membrane carries immobilized monoclonal antibodies to RSV fusion and internal proteins at the TEST line and goat anti-mou nonotibordu at the CONTROL line. The strip also contains colloidal gold conjugated to monoconnal anti-RSV fusion protein and anti-RSV internal protein as the detector antibodies. Sample Diluent: A buffered protein solution containing sodium azide (0.095%) as a preservative. Positive Control: Inactivated RSV in a buffered solution containing sodium azide (0.095%) as a preservative.
Here's a breakdown of the acceptance criteria and the study details for the ImmunoCard STAT! RSV device, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria are not explicitly stated as distinct numerical targets (e.g., "Sensitivity must be >85%"). Instead, the study aims to demonstrate substantial equivalence to predicate devices, which implies that the device's performance should be comparable to or better than the predicates. The performance metrics presented below are directly from the clinical trials that serve as the basis for proving equivalence.
| Performance Metric | Acceptance Criteria (Implied by Predicate Equivalence) | ImmunoCard STAT! RSV Reported Performance (Total Sites Combined, Resolved PCR Data) |
|---|---|---|
| Overall Agreement with Ground Truth | Comparable to predicate devices (e.g., Binax NOW RSV) | 92.1% (95% CI: 88.1% - 95.9%) |
| Sensitivity | Comparable to predicate devices | 85.9% (95% CI: 78.0% - 94.0%) |
| Specificity | Comparable to predicate devices | 96.3% (95% CI: 92.3% - 99.7%) |
| Positive Predictive Value (PPV) | Comparable to predicate devices | 93.8% (95% CI: 88.3% - 99.7%) |
| Negative Predictive Value (NPV) | Comparable to predicate devices | 91.2% (95% CI: 85.7% - 96.3%) |
2. Sample Size Used for the Test Set and Data Provenance
-
Test Set Sample Size:
- Total Samples (Combined Sites, unresolved): 174 (Table 6-5E)
- Total Samples (Combined Sites, resolved against PCR data at Site 2): 178 (Table 6-5F)
- Specific breakdown by sample type (resolved/unresolved not always distinct):
- Wash/Aspirate (unresolved): 48 (28 positive, 20 negative by culture)
- Wash/Aspirate (resolved): 52 (34 positive, 18 negative by culture)
- Swab (unresolved): 126 (37 positive, 89 negative by culture)
-
Data Provenance: Retrospective and prospective. Samples were collected from symptomatic patients.
- "archival (retrospective) or fresh (prospective) samples collected from symptomatic patients."
- The study involved three independent clinical laboratories, implying data from different geographical locations, though specific countries are not mentioned beyond "USA alone" in the background section regarding RSV hospitalization statistics.
3. Number of Experts Used to Establish Ground Truth and Qualifications
The document does not specify the number of individual experts or their specific qualifications (e.g., "radiologist with 10 years of experience") for establishing the primary ground truth.
- The ground truth methods used were tissue culture and PCR. These are laboratory-based diagnostic methods, not typically assessed by individual "experts" in the same way an image interpretation might be. The performance of these methods is inherent to the laboratory performing the assays, rather than a single individual's expertise. Clinical site 2 also performed PCR analysis for specimens with overgrowth in tissue culture or discrepant results, indicating a further analytical step by laboratory personnel.
4. Adjudication Method for the Test Set
- Primary Adjudication: Clinical Site 2 performed PCR analysis on specimens that exhibited overgrowth in tissue culture or showed discrepant results between the immunoassay and initial tissue culture. This PCR data was then used to "correct" or resolve the initial tissue culture ground truth for those specific samples. This acts as an adjudication step specifically for cases where the primary ground truth (tissue culture) was problematic or led to discordance.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- No, an MRMC comparative effectiveness study was not done. This device is a diagnostic immunoassay (a rapid test for antigen detection), not an imaging device or algorithm requiring human interpretation. Therefore, assessing how human readers improve with AI assistance is not applicable. The comparison is between the device's performance and established laboratory reference methods (tissue culture, PCR).
6. Standalone Performance Study (Algorithm Only)
- Yes, a standalone performance study was done. The entire clinical trial evaluates the ImmunoCard STAT! RSV device as a standalone diagnostic tool. Its performance (sensitivity, specificity, etc.) is directly compared against the established ground truth (tissue culture, sometimes resolved by PCR) without human interpretation as part of the device's output. The device itself produces a visual pink-red line for a positive result, and this is read directly.
7. Type of Ground Truth Used
- Tissue Culture: This was the primary diagnostic test standard for RSV detection.
- PCR (Polymerase Chain Reaction): Used as a secondary, confirmatory, or resolving ground truth for specific samples, particularly those with initial discrepant results or overgrowth in tissue culture.
8. Sample Size for the Training Set
- Not explicitly stated. The document describes clinical trials for performance evaluation (test set) but does not provide details on a specific training set size or methodology for the development of the ImmunoCard STAT! RSV assay. Immunoassays typically rely on antibody development and binding characteristics, rather than a machine learning "training set" in the computational sense. The "training" in this context would likely refer to the iterative development and optimization of the assay components and parameters.
9. How the Ground Truth for the Training Set Was Established
- Not explicitly described. As mentioned above, the concept of a "training set" with ground truth in the context of an immunoassay development differs from that of an AI algorithm. The development of the assay would have involved characterization of the monoclonal antibodies (directed at fusion and internal proteins of RSV) and optimization of the test device's chemical and physical properties to ensure reliable binding and signal generation. This process would inherently use characterized RSV samples (positive and negative) to establish optimal performance parameters, but these are not typically referred to as a "training set" with established "ground truth" in the same formal way as in AI/ML studies.
§ 866.3480 Respiratory syncytial virus serological reagents.
(a)
Identification. Respiratory syncytial virus serological reagents are devices that consist of antigens and antisera used in serological tests to identify antibodies to respiratory syncytial virus in serum. Additionally, some of these reagents consist of respiratory syncytial virus antisera conjugated with a fluorescent dye (immunofluorescent reagents) and used to identify respiratory syncytial viruses from clinical specimens or from tissue culture isolates derived from clinical specimens. The identification aids in the diagnosis of respiratory syncytial virus infections and provides epidemiological information on diseases caused by these viruses. Respiratory syncytial viruses cause a number of respiratory tract infections, including the common cold, pharyngitis, and infantile bronchopneumonia.(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.