(809 days)
The Aptiva APS IgG Reagent is an immunoassay utilizing particle-based multi-analyte technology for the semiquantitative determination of anti-cardiolipin (aCL) and anti-beta 2 glycoprotein 1 (all2GPI) IgG autoantibodies in human serum as an aid in the diagnosis of primary antiphospholipid syndrome (APS), when used in conjunction with other laboratory findings.
The Aptiva APS IgG Reagent is intended for use with the Aptiva System.
The Aptiva APS IgM Reagent is an immunoassay utilizing particle-based multi-analyte technology for the semiquantitative determination of anti-cardiolipin (aCL) and anti-beta 2 glycoprotein 1 (aß2GPI) IgM autoantibodies in human serum as an aid in the diagnosis of primary and secondary antiphospholipid syndrome (APS), when used in conjunction with other laboratory findings.
The Aptiva APS IgM Reagent is intended for use with the Aptiva System.
The Aptiva APS IgG and Aptiva APS IgM reagent utilize particle based multi-analyte technology (PMAT) in a cartridge format. Each analyte (anti-cardiolipin [aCL] and anti-B2-Glycoprotein I [aB2GPI]) in the Aptiva APS IgG and Aptiva APS IgM reagent is a solid phase immunoassay utilizing fluorescent microparticles. This technology allows each of the two analytes, along with a human IgG or human IgM capture antibody (IgG or IgM Control Microparticle), to be coated onto three uniquely recognizable paramagnetic microparticles, which are combined into one tube.
The Aptiva instrument is a fully automated, random-access analyzer. This platform is a closed system with continuous load and random-access capabilities that processes the samples, runs the reagent and reports results. It includes liquid handling hardware, optical module (OM), and integrated computer with proprietary software and touch screen user interface.
The two analyte microparticles, along with the control microparticle, are stored in the reagent cartridge under conditions that proteins in their reactive states. When the assay cartridge is ready to be used for the first time, the reagent tube seals are pierced using the cartridge lid. The reagent cartridge is then loaded onto the Aptiva instrument, where the microparticles are automatically rehydrated using a buffer located within the cartridge.
The Aptiva System dilutes the sample 1:8, then combines an aliquot of diluted sample, and reagent into a cuyette. The mixture is incubated at 37°C. After a wash cvcle, conjugated antihuman IgG or IdM antibodies are added to the particles and this mixture is incubated at 37°C. Excess conjugate is removed in another wash cycle, and the particles are re-suspended in system fluid.
Multiple images are generated by the system to identify and count the two (2) unique analyte particles, as well as determine the amount of coniugate on each particle. Coated with goat anti-human lgG or IdM antibodies, is present as a control to flaq low concentrations of IgG or IgM in the sample as an assay verification step. The median fluorescent intensity (MFI) for each analyte is proportional to the concentration of conjugate bound to human IgG or IgM, which is proportional to the concentration of IgG or IgM antibodies bound to the corresponding particle population. The system uses the MFI from at least 50 particles of each population. The identity of the particles is determined by the unique signature of the particles.
Each analyte in the Aptiva APS IgG Reagent and the Aptiva APS IgM Reagent is assigned a predefined lot specific master curve. The analyte specific master curve is stored on the reagent cartridge RFID label. Based on results obtained by running calibrators (supplied separately), the system creates individual working curves. Working curves are used by the software to calculate Fluorescent Light Units (FLU) for each analyte from the MFI values obtained for each sample.
Aptiva APS IgG and Aptiva APS IgM Calibrators and Aptiva APS IgG and Aptiva APS IgM Controls are sold separately.
The provided text describes the analytical and clinical performance characteristics of the Aptiva APS IgG and Aptiva APS IgM Reagents, which are immunoassays for the semi-quantitative determination of anti-cardiolipin (aCL) and anti-beta 2 glycoprotein 1 (aβ2GPI) IgG/IgM autoantibodies. This information is presented in the context of a 510(k) premarket notification for FDA clearance.
Here's a breakdown of the requested information based on the provided text:
1. A table of acceptance criteria and the reported device performance
The document does not explicitly state "acceptance criteria" as a separate, quantified set of thresholds for each performance metric. Instead, it presents the results of various analytical and clinical studies, implying that these results met the internal criteria for substantial equivalence to predicate devices and overall performance claims for an in vitro diagnostic (IVD) device.
However, we can infer performance targets based on the presented data and the overall context of an FDA submission for an IVD. The primary performance metrics presented are related to precision, detection limits, linearity, interference, and clinical sensitivity/specificity.
Inferred Acceptance Criteria and Reported Device Performance (Summary)
| Performance Characteristic | Inferred Acceptance Criteria (General IVD Expectations) | Reported Device Performance (Aptiva APS Reagents) |
|---|---|---|
| Precision | CV% to be within acceptable ranges for IVD assays, typically lower for higher concentrations and clinically critical ranges. | Within-Laboratory (Total Precision) CV%: |
- aCL IgG: 5.6% - 9.5% (generally decreasing with higher FLU)
- aβ2GPI IgG: 6.9% - 11.7% (generally decreasing with higher FLU)
- aCL IgM: 4.3% - 9.7% (generally decreasing with higher FLU)
- aβ2GPI IgM: 5.5% - 10.2% (generally decreasing with higher FLU)
Between-Site Reproducibility CV% (3 sites):
- aCL IgG: 5.2% - 9.3%
- aβ2GPI IgG: 6.4% - 10.0%
- aCL IgM: 5.4% - 10.0%
- aβ2GPI IgM: 5.9% - 10.5%
Between-Lot Reproducibility CV% (3 lots):
- aCL IgG: 6.6% - 13.3%
- aβ2GPI IgG: 8.5% - 12.1%
- aCL IgM: 6.1% - 11.4%
- aβ2GPI IgM: 6.0% - 10.5% |
| Limit of Blank (LoB) | Very low, close to zero, ensuring no signal from blank samples. | aCL IgG: 0.00 FLU
aβ2GPI IgG: 0.02 FLU
aCL IgM: 0.01 FLU
aβ2GPI IgM: 0.03 FLU |
| Limit of Detection (LoD)| Low, indicating ability to detect small amounts of analyte. | aCL IgG: 0.07 FLU
aβ2GPI IgG: 0.09 FLU
aCL IgM: 0.04 FLU
aβ2GPI IgM: 0.06 FLU |
| Limit of Quantitation (LoQ)| Low, defining the lowest concentration that can be reliably quantified. | aCL IgG: 0.29 FLU
aβ2GPI IgG: 0.21 FLU
aCL IgM: 0.06 FLU (set to 0.10 FLU for AMR lower limit)
aβ2GPI IgM: 0.09 FLU (set to 0.10 FLU for AMR lower limit) |
| Analytical Measuring Range (AMR)| Wide enough to cover relevant clinical concentrations, with demonstrated linearity. | aCL IgG: 0.29 - 328.94 FLU
aβ2GPI IgG: 0.21 - 256.70 FLU
aCL IgM: 0.10 – 114.68 FLU
aβ2GPI IgM: 0.10 – 95.86 FLU
Linearity demonstrated across these ranges with R2 values mostly ≥ 0.98. |
| High Concentration Hook Effect| No hook effect within or above the AMR. | Confirmed no hook effect up to theoretically calculated values: aCL IgG: 2645.36 FLU, aβ2GPI IgG: 1790.48 FLU, aCL IgM: 167.25 FLU, aβ2GPI IgM: 126.13 FLU. |
| Interference | No significant interference from common endogenous or exogenous substances at specified concentrations. | No interference detected for aCL IgG, aβ2GPI IgG, aCL IgM, and aβ2GPI IgM with tested interferents (bilirubin, hemoglobin, triglycerides, cholesterol, RF IgM, human IgG, ibuprofen, warfarin, prednisone, acetaminophen, aspirin, hydroxychloroquine, omeprazole, simvastatin, heparin) at their respective tested concentrations. Percent recoveries or FLU differences were within acceptable ranges (generally close to 100% recovery for spiked samples, or low FLU difference for negative samples). |
| Sample Stability | Samples should be stable for specific storage conditions and freeze/thaw cycles. | Samples stable up to 48 hours at room temperature, up to 14 days at 2-8°C, and for up to 5 freeze/thaw cycles. |
| Reagent Stability | Reagent shelf-life and in-use stability should be established. | Shelf-life: 9 months for Aptiva APS IgG Reagent, 7 months for Aptiva APS IgM Reagent (based on accelerated stability, verified by ongoing real-time studies).
In-use (onboard) stability: 28 days for both, with 14-day recalibration. |
| Clinical Sensitivity & Specificity| High sensitivity to detect disease (APS) and high specificity to correctly identify non-disease states (controls/non-APS). | Aptiva APS IgG:
- aCL IgG: Sensitivity 54.1% (95% CI: 45.3–62.7%), Specificity 99.5% (95% CI: 98.2–99.9%)
- aβ2GPI IgG: Sensitivity 53.3% (95% CI: 44.5-61.9%), Specificity 99.0% (95% CI: 97.5-99.6%)
Aptiva APS IgM:
- aCL IgM: Sensitivity 27.5% (95% CI: 22.7–32.9%), Specificity 97.5% (95% CI: 95.4–98.6%)
- aβ2GPI IgM: Sensitivity 24.7% (95% CI: 20.1–30.0%), Specificity 98.5% (95% CI: 96.8–99.3%) |
| Predicate Method Comparison (Percent Agreement)| High agreement with legally marketed predicate devices. | Aptiva APS IgG (aCL IgG) vs. QUANTA Flash aCL IgG: PPA: 81.6%, NPA: 95.7%, TPA: 93.1% (N=202)
Aptiva APS IgG (aβ2GPI IgG) vs. QUANTA Lite Beta 2GP1 IgG ELISA: PPA: 88.0%, NPA: 89.7%, TPA: 88.9% (N=108)
Aptiva APS IgM (aCL IgM) vs. QUANTA Flash aCL IgM: PPA: 87.0%, NPA: 90.2%, TPA: 89.8% (N=422)
Aptiva APS IgM (aβ2GPI IgM) vs. QUANTA Flash β2GPI IgM: PPA: 88.9%, NPA: 84.3%, TPA: 84.8% (N=244) |
2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
-
Clinical Performance Test Set Sample Sizes:
- Aptiva APS IgG (aCL IgG & aβ2GPI IgG): N=526 (122 APS combined, 404 controls/non-APS)
- Aptiva APS IgM (aCL IgM & aβ2GPI IgM): N=689 (291 APS combined, 398 controls/non-APS)
- Normal Population for Expected Values: N=200 apparently healthy blood donors.
-
Method Comparison Test Set Sample Sizes:
- Aptiva APS IgG (aCL IgG vs. QUANTA Flash aCL IgG): N=202
- Aptiva APS IgG (aβ2GPI IgG vs. QUANTA Lite Beta 2GP1 IgG ELISA): N=108
- Aptiva APS IgM (aCL IgM vs. QUANTA Flash aCL IgM): N=422
- Aptiva APS IgM (aβ2GPI IgM vs. QUANTA Flash β2GPI IgM): N=244
-
Analytical Performance Test Set Sample Sizes:
- Precision: 7 samples for IgG, 7 samples for IgM (80 replicates each for within-lab; 75 replicates each for between-site/lot reproducibility from multiple sites/lots).
- LoB/LoD/LoQ: Blanks (LoB: 4 samples, 60 data points/lot); Low-level samples (LoD/LoQ: 4 samples, 120 data points/assay/lot).
- Interference: 6 human specimens (negative, cutoff, positive) for each analyte, spiked with various interferents and tested in 5 replicates.
- Sample Stability: 5 serum samples (IgG), 5 serum samples (IgM) tested in duplicates over time/cycles.
- In-use Stability: 11 samples (IgG), 7 samples (IgM) tested periodically.
-
Data Provenance: The document states that a "cohort of characterized samples, none of which were used for establishing the reference range, was used to validate the clinical performance." It does not explicitly state the country of origin of the data or whether the studies were retrospective or prospective. However, for a 510(k) submission, clinical validation studies typically involve retrospective or prospectively collected clinical samples, but the exact nature (e.g., specific clinical sites, patient populations beyond disease groups) and geographic origin are not detailed here. The studies were likely conducted within a controlled laboratory setting by the manufacturer (Inova Diagnostics, Inc. in San Diego, CA) or its affiliates, using sourced human serum samples.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)
The establishment of "ground truth" for IVD devices like these typically relies on well-characterized clinical samples and established diagnostic criteria for the disease (Antiphospholipid Syndrome - APS).
- The document states that the clinical performance validation was performed using "a cohort of characterized samples." The characterization of these samples (i.e., whether they definitively represent APS or control) would serve as the ground truth.
- However, the document does not specify the number of experts or their qualifications (e.g., rheumatologists, clinical immunologists/pathologists) who established the diagnostic status (ground truth) of the clinical samples (APS vs. control) used in the clinical sensitivity and specificity studies. It is implied that these were "characterized samples," meaning their disease status was determined by established clinical and laboratory criteria, likely involving clinical consensus or previous diagnoses.
- For cut-off establishment, the reference population included "apparently healthy subjects," and the "internal APS samples (data not provided)" and "distribution of result values of healthy controls" were used. This suggests clinical characterization of these samples.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
The concept of "adjudication method" (like 2+1 or 3+1) is typically relevant for interpretative tasks, such as reading medical images, where multiple human readers interpret the same data and their interpretations need to be reconciled to establish a ground truth.
For these types of IVD assays, ground truth for clinical performance is established based on the clinical diagnosis of the patient from whom the sample was collected. This diagnosis is usually a culmination of clinical findings, established criteria (e.g., the revised Sapporo criteria for APS), and other laboratory tests, rather than an "adjudication" of multiple independent interpretations of the test results themselves.
Therefore, the document does not mention any adjudication method in the context of establishing ground truth for the test samples, as it's not applicable in the same way it would be for an AI-medical imaging device. The "ground truth" for the samples (APS vs. non-APS) is assumed to be pre-established clinical diagnosis.
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 MRMC study was conducted or is applicable here.
This device is an in vitro diagnostic (IVD) immunoassay, not an AI-powered image analysis or diagnostic aid that assists human readers (e.g., radiologists interpreting images). The device directly measures biomarker levels in a sample, and its output is a quantitative value (FLU) which then determines a semi-quantitative result (Positive/Negative/Indeterminate based on cut-offs). Human "readers" (laboratory personnel) operate the instrument and interpret the final quantitative results based on predefined cut-offs, but they are not subjectively interpreting complex data that AI would assist with, in the sense of an MRMC study.
Therefore, an MRMC comparative effectiveness study, and an effect size related to human reader improvement with AI assistance, are not relevant for this type of device and are not mentioned in the document.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
This is a "standalone" device in terms of its core functionality, but the term "algorithm only" or "human-in-the-loop" isn't directly analogous.
- The Aptiva System is a fully automated, random-access analyzer (page 6). This means the instrument itself, with its integrated software and optical module, processes the samples, runs the reagents, and reports results independently after the sample is loaded and the assay initiated.
- The "performance" described here (sensitivity, specificity, precision, linearity, etc.) is the device's performance (including its internal algorithms and mechanics) in generating quantitative results. There isn't a separate "algorithm only" performance that needs to be differentiated from a "human-in-the-loop" performance, because the device is the automated system determining the FLU values. The human interaction is primarily in sample loading, reagent handling, and result review/reporting, not in interpreting raw data that the device itself would also interpret in an unassisted mode.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)
The ground truth for the clinical performance studies (sensitivity and specificity) was established based on "characterized samples" representing patients with Antiphospholipid Syndrome (APS) and various control groups (patients with other autoimmune/infectious diseases, apparently healthy subjects).
While the document doesn't explicitly state "expert consensus," it's highly implied that the "characterization" of these samples as APS or control would be based on:
- Clinical findings: Presenting symptoms, medical history.
- Other laboratory findings: Beyond the novel antibodies, other relevant diagnostic tests.
- Established diagnostic criteria: For APS, this would typically be the revised Sapporo classification criteria, which combine clinical and laboratory criteria.
So, it's a combination of established clinical diagnoses and potentially other laboratory data, which implicitly would involve the consensus or findings of medical experts involved in patient diagnosis. It is not based on pathology (e.g., tissue biopsy) or outcomes data (e.g., long-term disease progression as the sole ground truth).
8. The sample size for the training set
The document describes the submission of a "new device" and its performance characteristics. It does not explicitly mention or quantify a "training set" in the context of machine learning.
For IVD devices, a "training set" isn't a standard concept unless the device incorporates adaptive algorithms or AI that learns from data. In this case, the device is an immunoassay with predefined master curves and calibrated reagents. The master curves are generated "at Inova for each reagent lot, where in-house Master Curve Standards with assigned FLU values are run multiple times." These "in-house Master Curve Standards" could be considered analogous to a "training" or calibration process, but it's not a dataset for training a generalized AI model but rather for calibrating each reagent lot of a classical assay.
The sample sizes provided in the document are for:
- Analytical performance (precision, LoB/LoD/LoQ, linearity, interference, stability).
- Clinical validation (sensitivity/specificity studies).
- Method comparison studies.
- Reference range establishment.
None of these are explicitly labeled as a "training set."
9. How the ground truth for the training set was established
As there is no explicitly defined "training set" for a machine learning model, the concept of establishing ground truth for such a set is not applicable here.
However, if we consider the "Master Curve Standards" as analogous to calibration/training data for the assay, their "ground truth" (assigned FLU values) would be established through a rigorous internal process by the manufacturer (Inova Diagnostics) based on:
- Carefully prepared and characterized aliquots (standards) with known or assigned concentrations of the target antibodies.
- Repeat measurements and statistical analysis for consistent and accurate assignment of FLU values.
- This is a standard practice for calibrating quantitative IVD assays, ensuring the device outputs accurate and traceable results.
§ 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).