Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K190675
    Date Cleared
    2019-05-17

    (63 days)

    Product Code
    Regulation Number
    862.1215
    Reference & Predicate Devices
    Predicate For
    N/A
    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Dimension EXL High-Sensitivity Troponin I (TNIH) assay is for in vitro diagnostic use in the quantitative measurement of cardiac troponin I in human plasma using the Dimension EXL integrated chemistry system with LOCI module. The assay can be used to aid in the diagnosis of acute myocardial infarction (AMI).

    Device Description

    The Dimension EXL TNIH assay is a homogeneous, sandwich chemiluminescent immunoassay based on LOCl® technology. The LOCI reagents include two synthetic bead reagents and two biotinylated anti-cardiac troponin I monoclonal antibody fragments. The first bead reagent (Sensibeads) is coated with streptavidin and contains photosensitizer dye. The second bead reagent (Chemibeads) is coated with a third anticardiac troponin I monoclonal antibody and contains chemilyminescent dve. Sample is incubated with Chemibeads and biotinylated antibodies to form bead-cardiac troponin Ibiotinylated antibody sandwiches. Sensibeads are added and bind to form bead-pair immunocomplexes. Illumination of the complex at 680 nm generates singlet oxygen from Sensibeads which diffuses into the Chemibeads, triggering a chemiluminescent reaction. The resulting signal is measured at 612 nm and is a direct function of the cardiac troponin I concentration in the sample. Lithium heparin plasma specimens may be used. The reagent is stored unopened at 2 -8 °C, is stable sealed on system for 30 days and opened on the system for 7 days. Calibration is performed every 21 days for a reagent lot.

    AI/ML Overview

    Here's a summary of the acceptance criteria and study details for the Dimension EXL High-Sensitivity Troponin I (TNIH) Assay, extracted from the provided text:

    1. Table of Acceptance Criteria and Reported Device Performance

    Performance MetricAcceptance Criteria (implicit from compliance)Reported Device Performance
    Limit of Blank (LoB)Consistency with claim of 1.1 pg/mL1.0 pg/mL
    Limit of Detection (LoD)Consistency with claim of 2.7 pg/mL1.0 - 1.8 pg/mL
    Limit of Quantitation (LoQ)CV ≤ 20.0%1.1 pg/mL (Lot-1), 2.1 pg/mL (Lot-2), 2.3 pg/mL (Lot-3). Claimed LoQ of 4.0 pg/mL. (The LoQ of 4.0 pg/mL is explicitly stated to be consistent with the data, implying it met the 20% CV criteria).
    10% CV LimitCV ≤ 10.0%3.2 pg/mL (Lot-1), 5.3 pg/mL (Lot-2), 5.8 pg/mL (Lot-3). Claimed 10% CV limit of 12.0 pg/mL. (The 10% CV limit of 12.0 pg/mL is explicitly stated to be consistent with the data, implying it met the 10% CV criteria).
    Precision (Repeatability %CV)All precision goals were met (specific target not provided, but values shown are low)Plasma 1: 2.3%, Plasma 2: 2.0%, Plasma 3: 1.8%, QC: 2.0%
    Precision (Within-Lab %CV)All precision goals were met (specific target not provided, but values shown are low)Plasma 1: 6.0%, Plasma 2: 2.9%, Plasma 3: 3.0%, QC: 3.3%
    Linearityp-values of nonlinear terms > 0.05 OR allowable bias ≤ 10% or 4 pg/mLConfirmed linearity from 4.0 - 25,000.0 pg/mL.
    Interferences (Bias)Bias exceeding 10% is considered interference.Hemoglobin: <10%, Bilirubin (conjugated): <10%, Bilirubin (unconjugated): <10%, Lipemia: <10%. No interference from a wide range of therapeutic/toxic substances (listed in detail, all <10%).
    Cross-reactivityLower values indicate less cross-reactivity (no explicit threshold provided, but shown to be very low)Cardiac Troponin T: 0.003%, Skeletal Troponin I: 0.001%, Tropomyosin: ND, Actin: ND, Troponin C: ND, Myosin Light Chain: ND, Myoglobin: ND, CK-MB: ND.
    High Dose Hook EffectNo hook effectNo hook effect found at 1,000,000 pg/mL troponin.
    Dilution RecoveryTesting supported use of diluent for over-range samples.Supported use of CTNI SDIL as diluent for 1:2 and 1:5 dilutions of native human plasma samples with high TNIH levels.
    Calibration Stabilityp-value of regression slope ≥ 0.05 OR drift ≤ LoQ or ≤ 10% (up to 20,000 pg/mL), ≤ 13% (> 20,000 pg/mL)Calibration interval was measured to be 21 days. (This implies it met the acceptance criteria for 21 days).
    Open Well Stabilityp-value of regression slope ≥ 0.05 OR drift ≤ LoQ or ≤ 10% (up to 20,000 pg/mL), ≤ 13% (> 20,000 pg/mL)Stability of reagents opened onboard the instrument was 7 days per well set. (This implies it met the acceptance criteria for 7 days).
    Sample StabilityLower bound of 95% CI of regression line ≤ -10% AND all data points bias ≤ -20% vs time zeroSeparated samples stable for 8 hours at room temperature, 24 hours at 2-8°C, up to 40 days at or below -20°C (non-frost free), and up to 1 year at or below -70°C. (This implies these storage conditions met the specified bias criteria).
    99th Percentile Expected ValuesDetermined using non-parametric statistical method (CLSI EP28-A3c).Overall: 60.4 pg/mL (90% CI: 43.2 to 81.3); Female: 51.4 pg/mL (90% CI: 35.6 to 109.2); Male: 76.2 pg/mL (90% CI: 42.3 to 117.0). No statistical basis for using separate values based on gender or sample type.
    Clinical PerformanceNo explicit quantitative acceptance criteria provided in the extract, but the data presented supports its diagnostic utility.Pooled Gender (60.4 pg/mL cutoff): Sensitivity ranging from 78.3% (0-<1.5h) to 95.5% (4.5-<6h), Specificity ranging from 86.2% (≥9-<24h) to 92.9% (0-<1.5h). NPV consistently high (e.g., 96.8% at 0-<1.5h, up to 99.3% at 4.5-<6h).Female-specific (51.4 pg/mL cutoff): Sensitivity ranging from 83.3% (0-<1.5h) to 100.0% (≥24h), Specificity ranging from 82.0% (≥24h) to 94.6% (0-<1.5h). NPV consistently high (e.g., 98.2% at 0-<1.5h, up to 99.5% at 4.5-<6h).Male-specific (76.2 pg/mL cutoff): Sensitivity ranging from 74.0% (0-<1.5h) to 90.9% (≥9-<24h), Specificity ranging from 87.2% (≥9-<24h) to 92.7% (≥3.5-<4.5h). NPV consistently high (e.g., 95.5% at 0-1.5h, up to 97.7% at 3.5-<4.5h).

    2. Sample size used for the test set and the data provenance

    • Clinical Performance (diagnostic accuracy):
      • Sample Size: Approximately 2500 subjects.
      • Data Provenance: Prospective study, specimens collected at 29 emergency departments across the United States from subjects presenting with symptoms consistent with acute coronary syndrome (ACS).
    • Expected Values (99th percentile):
      • Sample Size: 2020 apparently healthy individuals (1017 female, 1003 male).
      • Data Provenance: Lithium heparin plasma specimens collected from apparently healthy individuals from the United States.

    3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

    • Number of Experts: Panels of certified cardiologists and emergency physicians (specific number not given, but implied to be multiple for "panels").
    • Qualifications: Certified cardiologists and emergency physicians.

    4. Adjudication method for the test set

    • Adjudication Method: "All subject diagnoses were adjudicated by panels of certified cardiologists and emergency physicians according to the Third Universal Definition Of Myocardial Infarction - consensus guideline endorsed by the European Society of Cardiology (ESC), the American College of Cardiology Foundation (ACCF), the American Heart Association (AHA), and the World Heart Federation (WHF)." This indicates a consensus-based adjudication following established clinical guidelines. Specific numerical adjudication (e.g., 2+1) is not explicitly stated.

    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 comparative effectiveness study was done, nor is this applicable as it is an in vitro diagnostic assay, not an AI-assisted imaging device or system. The study focuses on the standalone performance of the assay compared to a clinically adjudicated diagnosis.

    6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

    • Yes, a standalone study was done. The entire performance evaluation here describes the analytical and clinical performance of the assay itself (the Dimension EXL TNIH assay) as a direct measurement tool, without any human-in-the-loop interpretation or AI assistance. The clinical performance section explicitly evaluates the assay's sensitivity, specificity, PPV, and NPV against the adjudicated diagnoses.

    7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

    • Ground Truth Type: Expert consensus based on a recognized international clinical guideline (Third Universal Definition Of Myocardial Infarction endorsed by various cardiology and heart associations).

    8. The sample size for the training set

    • The document does not explicitly mention a "training set" in the context of an algorithm or AI. This is a traditional immunoassay, and its development would typically involve internal optimization and validation, not a distinct "training set" like that for machine learning. The data presented here describes the performance evaluation during regulatory submission.

    9. How the ground truth for the training set was established

    • Not applicable, as a distinct training set with established ground truth, as understood in the context of AI/ML, is not described for this immunoassay.
    Ask a Question

    Ask a specific question about this device

    Page 1 of 1