(273 days)
The Gold Standard Diagnostics AIX1000 Rapid Plasma Reagin (RPR) Automated Test System is a non-treponemal flocculation test that can qualitatively determine the presence of reagin antibodies in human serum. It may be used to aid in the diagnosis of syphilis when used in conjunction with supplemental treponemal laboratory tests and other clinical information. This test may also be used to detect non-treponemal antibodies in samples serially diluted to establish titer information. This test is not intended for screening blood or tissue donors.
The Gold Standard Diagnostics AIX1000 Rapid Plasma Reagin (RPR) Automated Test System in a non-treponemal test for the qualitative determination of reagin antibodies in human serum to aid in the diagnosis of syphilis. This test is also used to detect non-treponemal antibodies in samples serially diluted to establish titer information. The system consists of the Gold Standard Diagnostics RPR test reagents and the Gold Standard Diagnostics AIX1000 Agglutination Analyzer. The AIX1000 Analyzer delivers serum from collection tubes into test wells. After antigen suspension is added, the test wells are then incubated while being shaken. An onboard camera creates a high resolution image. The image is then analyzed by the proprietary software algorithm to produce a result.
Here's an analysis of the acceptance criteria and the studies that prove the device meets these criteria, based on the provided text:
Device Name: Gold Standard Diagnostics AIX1000 Rapid Plasma Reagin (RPR) Automated Test System
K Number: K150358
1. Table of Acceptance Criteria and Reported Device Performance
The document doesn't explicitly state "acceptance criteria" in a bulleted or numbered list with predefined thresholds for all studies. However, performance expectations are implied by the nature of the tests conducted and the reported agreement percentages. I will interpret the reported performance metrics as demonstrating the device meets an implied acceptance for its intended use, particularly through comparison to a predicate device.
| Study Type / Performance Metric | Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|---|
| Cross Reactivity | No observed cross-reactivity with common viral, bacterial, and autoimmune conditions. | Viral: Rubella (10), VZV (10), HIV (10), Hepatitis B (16), Hepatitis C (11), EBV (10), HSV Type 1 (10), HSV Type 2 (10), CMV (11), Heterophile antibodies (10) - All 0 Reactive. |
| Bacterial: Toxoplasma gondii (10), Leptospira biflexa (10), Borrelia burgdorferi (10) - All 0 Reactive. | ||
| Autoimmune: SLE (10), Rheumatoid Arthritis (10), Scleroderma (10), Primary Anti-phospholipid Syndrome (16) - All 0 Reactive. (Numbers in parentheses are 'Number Tested'). | ||
| Interfering Substances | No effect on performance by specified endogenous substances and prescription drugs. | Hemoglobin (20 g/dL), Bilirubin (unconjugated) (15 mg/dL), Cholesterol (250 mg/dL), Albumin (5 g/dL), Gamma Globulin (60 mg/dL), Glucose (120 mg/dL), Triglyceride (500 mg/dL), Antibiotic (Cephalexin) (337 umol/L), Antibiotic (Tetracycline) (34 umol/L). For all substances, RPR samples remained positive where expected; "None Observed" for interference. |
| Precision (Within-Lab) | High agreement (e.g., >90%) within ± 1 titer for various reactivity levels. | - Non-Reactive Serum: 100% agreement (93.6% - 100% C.I.) |
- Low RPR Reactivity (1:4): 100% agreement (93.6% - 100% C.I.)
- Moderately Reactive (1:16): 100% agreement (93.6% - 100% C.I.)
- Reactive (1:64): 97.8% agreement (88.2% - 99.9% C.I.)
- Highly Reactive (1:128): 100% agreement (93.6% - 100% C.I.)
- Reactive control: 100% agreement (54.9% - 100% C.I.)
- Non-reactive control: 100% agreement (54.9% - 100% C.I.) |
| Reproducibility | High agreement (e.g., >90%) within ± 1 titer across operators, instruments, days, and runs. | - Non-Reactive Serum: 100% (Between-Runs, Between-Days, Between-Operators, Between-Instruments) - Low RPR Reactivity (1:4): 100% (all categories)
- Moderately Reactive (1:16): 100% (all categories)
- Reactive (1:64): 100% (Between-Runs, Between-Operators, Between-Instruments), 97.8% (Between-Days)
- Highly Reactive (1:128): 100% (Between-Runs, Between-Days, Between-Instruments), 98.1% (Between-Operators)
- Reactive control: 100% (90.5% - 100% C.I.)
- Non-reactive control: 100% (90.5% - 100% C.I.) |
| Carry-over | No evidence of carry-over (no false positives in negative samples due to highly reactive samples). | All 480 replicates of the negative samples were reported as non-reactive when highly reactive samples were alternated with non-reactive samples. |
| Clinical Agreement (Prospective Samples vs. Predicate) | High Positive Percent Agreement (PPA) and Negative Percent Agreement (NPA) with predicate. | PPA: 95.5% (C.I. 77.2% - 99.9%)
NPA: 99.9% (C.I. 99.3% - 100%) |
| Clinical Agreement (Retrospective Samples vs. Predicate) | High PPA and NPA with predicate. | PPA: 97.2% (C.I. 95.5% - 98.4%)
NPA: 99.1% (C.I. 98.5% - 99.5%) |
| Special Populations (Pregnant Women vs. Predicate) | High PPA and NPA with predicate. | PPA: 100% (C.I. 90.5% - 100%)
NPA: 100% (C.I. 98.8% - 100%) |
| Special Populations (HIV Positive Individuals vs. Predicate) | High PPA and NPA with predicate. | PPA: 100% (C.I. 90.5% - 100%)
NPA: 100% (C.I. 98.8% - 100%) |
| Healthy Individuals | Low and expected reactivity in a healthy, no-risk population. | All 100 samples from apparently healthy individuals (not at risk, no syphilis test ordered) were non-reactive. |
| Clinical Stages Correlation | High agreement with clinically diagnosed syphilis cases across various stages. | Primary Treated (13/13 reactive), Primary Untreated (12/12 reactive), Secondary Treated (25/25 reactive), Secondary Untreated (25/25 reactive), Latent Treated (25/25 reactive), Latent Untreated (25/25 reactive). All showed 100% agreement with corresponding 95% C.I. ranges of 77.9%-100% to 88.7%-100%. |
2. Sample Sizes Used for the Test Set and Data Provenance
The document describes several test sets:
- Cross Reactivity: Panels of 10-16 individual patient samples per condition (totaling 17 conditions for viral, bacterial, and autoimmune).
- Interfering Substances: 5 samples (1 non-reactive, 4 reactive) per interfering substance.
- Precision: 5 samples (Non-reactive, Low RPR (1:4), Moderately Reactive (1:16), Reactive (1:64), Highly Reactive (1:128)). Each tested in 9 replicates.
- Reproducibility: The same 5 sample panels as precision, tested across 6 operators, 3 instruments, 2 runs over 5 days.
- Carry-over: 1 reactive (1:64), 1 highly reactive (1:128), and 2 non-reactive samples. Highly reactive samples alternated with non-reactive samples 96 times per run over 5 runs.
- Clinical Studies - Prospectively Collected Samples:
- Sample Size: 765 serum samples.
- Data Provenance: Collected prospectively from patient samples with a physician's order for syphilis testing. Collected at two geographically distinct reference laboratories (Southeastern and Western United States).
- Clinical Studies - Retrospectively Collected Samples:
- Sample Size: 2,246 samples.
- Data Provenance: Collected retrospectively from patients referred for syphilis testing. Obtained from sample brokers who collect from multiple sites across the United States. Collection dates: January 2005 - July 2014.
- Clinical Studies - Pregnant Women (Retrospective):
- Sample Size: 250 non-reactive, 30 reactive (spiked) samples. Total 280.
- Data Provenance: Retrospectively collected from pregnant women at one site (Southeastern United States). Collection dates: July 2012 - August 2013.
- Clinical Studies - HIV Positive Individuals (Retrospective):
- Sample Size: 250 non-reactive, 30 reactive samples. Total 280.
- Data Provenance: Retrospectively collected from HIV positive individuals at four sites (one Southeastern, one Mid-Western States). Collection dates: February 2012 - June 2015.
- Apparently Healthy Individuals (Prospective):
- Sample Size: 100 serum samples.
- Data Provenance: Prospectively collected from healthy individuals not at risk for syphilis and for whom a syphilis test had not been ordered (submitted for routine chemistry testing).
- Correlation with Clinically Diagnosed Syphilis Sera - Various Stages (Purchased Panel):
- Sample Size: Primary Treated (13), Primary Untreated (12), Secondary Treated (25), Secondary Untreated (25), Latent Treated (25), Latent Untreated (25).
- Data Provenance: Purchased from serum brokers, characterized by clinical diagnosis.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
The document does not mention the use of human experts to establish ground truth for the test sets in the context of adjudication or interpretation of the RPR results. Instead, it relies on:
- Reference laboratory results (for prospective samples).
- Predicate device results (for comparison in clinical studies).
- FDA-cleared treponemal (TP) assays (for further investigation of discrepant non-treponemal results).
- Clinical diagnosis (for the purchased syphilis stage panel).
- Serum broker confirmation of disease markers (for cross-reactivity samples).
Therefore, there is no direct information on the number or qualifications of experts involved in establishing ground truth through manual interpretation of RPR images or similar tasks for the "AIX1000" aspect. The 'AI' component is described as proprietary software interpreting images (standalone).
4. Adjudication Method for the Test Set
Adjudication methods (like 2+1, 3+1) are typically used when multiple human readers interpret the same image/case and disagreements need to be resolved. This device is an automated test system for a flocculation assay where the "AIX1000 Agglutination Analyzer" uses a proprietary software algorithm to produce a result.
For the discrepancies in the clinical studies where the GSD AIX1000 RPR Test System disagreed with the comparator device:
- Prospective Samples: The two discrepant samples were tested on a "third FDA cleared RPR assay."
- Retrospective Samples: The 31 discrepant samples were tested on a "third FDA cleared RPR assay."
This suggests a form of tie-breaking or external reference testing for discrepant results, rather than a human expert adjudication of visual interpretations by the device itself.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was Done
No, an MRMC comparative effectiveness study was not conducted. The device is an automated test system that interprets RPR assay results without direct human intervention in the interpretation process described. The comparisons were between the automated device and a predicate automated device or clinical outcomes, not between human readers with and without AI assistance.
6. If a Standalone Performance (i.e., algorithm only without human-in-the-loop performance) was Done
Yes, the studies reported represent standalone performance of the Gold Standard Diagnostics AIX1000 RPR Automated Test System. The device's proprietary software algorithm directly produces the results from images captured by its onboard camera. The clinical studies (prospective, retrospective, special populations) directly compare the device's output to a predicate device's output or clinical diagnosis, without human modification or override of the AIX1000's initial determination. The interpretation of the flocculation assay itself is automated.
7. The Type of Ground Truth Used
The ground truth or reference methods varied depending on the study:
- Clinical Studies (Prospective and Retrospective): The primary ground truth for comparison was the legally marketed predicate device (Arlington Scientific Inc. (ASI) RPR Card Test for syphilis on the ASiManager-AT Analyzer). For discrepant samples, a third FDA cleared RPR assay was used as a tie-breaker. Further investigation for NT+ samples used an FDA cleared treponemal (TP) assay.
- Correlation with Clinically Diagnosed Syphilis Sera: Clinical Diagnosis (Primary, Secondary, Latent stages as characterized by documented symptoms, dark field microscopy, and reactive treponemal/non-treponemal tests).
- Cross-Reactivity Study: Confirmation of disease markers by serum brokers.
- Special Populations (Pregnant Women, HIV Positive): Predicate device results, and also confirmed pregnancy status via HCG test and HIV positive status.
- Apparently Healthy Individuals: Lack of risk factors and no physician's order for syphilis testing defined this as "healthy."
8. The Sample Size for the Training Set
The document does not provide information on the sample size used for the training set for the proprietary software algorithm. This information is typically proprietary to the manufacturer and not usually detailed in 510(k) summaries unless specifically requested or deemed critical for substantial equivalence. The focus of the 510(k) is often on comparing the new device's performance to a predicate, rather than the internal development of its algorithm.
9. How the Ground Truth for the Training Set Was Established
Similar to the training set size, the document does not provide information on how the ground truth for the training set was established. Since the device uses a "proprietary software algorithm" to analyze images for flocculation, it can be inferred that this algorithm was developed and trained using a collection of RPR test results with established reactivity and titer values. However, the specifics of this ground truth (e.g., expert consensus on visual flocculation, correlation with other assays) are not detailed in this submission.
§ 866.3820
Treponema pallidum nontreponemal test reagents.(a)
Identification. Treponema pallidum nontreponemal test reagents are devices that consist of antigens derived from nontreponemal sources (sources not directly associated with treponemal organisms) and control sera (standardized sera with which test results are compared) used in serological tests to identify reagin, an antibody-like agent, which is produced from the reaction of treponema microorganisms with body tissues. The identification aids in the diagnosis of syphilis caused by microorganisms belonging to the genusTreponema and provides epidemiological information on syphilis.(b)
Classification. Class II (performance standards).