LogoFDA 510(k) Search
Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K171642
    Device Name
    Atellica IM Ferritin Assay
    Manufacturer
    Siemens Healthcare Diagnostics Inc.
    Date Cleared
    2017-08-31

    (90 days)

    Product Code
    DBF
    Regulation Number
    866.5340
    Type
    Traditional
    Panel
    Immunology
    Reference & Predicate Devices
    k041133,k992157
    Predicate For
    N/A
    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Atellica™ IM Ferritin (Fer) assay is for in vitro diagnostic use in the quantitative determination of ferritin in human serum and plasma (EDTA and lithium heparin) using the Atellica™ IM Analyzer. This assay can be used as an aid in the diagnosis of iron deficiency anemia and iron overload.

    Device Description

    The Atellica Ferritin Assay kit includes the following components: Lite Reagent: 5.0 mL/reagent pack. Contains Goat polyclonal anti-ferritin antibody (~0.64 µg/mL) labeled with acridinium ester in HEPES buffer; protein stabilizers; sodium azide (< 0.1%); preservatives Solid Phase Reagent: 22.5 mL/reagent pack. Contains Mouse monoclonal anti-ferritin antibody (~32.2 µg/mL) covalently coupled to paramagnetic particles in sodium barbital buffer; protein stabilizers; sodium azide (< 0.1%); preservatives

    AI/ML Overview

    Here's a breakdown of the acceptance criteria and study details for the Atellica IM Ferritin Assay, based on the provided document:

    Acceptance Criteria and Reported Device Performance

    CriteriaAcceptance CriteriaReported Device Performance
    PrecisionCLSI EP05-A3 guidelines (Evaluation of Precision Performance of Quantitative Measurement Methods)Repeatability (Within-run): CV ranges from 1.2% to 3.5% for samples (4.2 ng/mL to 1453.6 ng/mL) and 1.2% to 1.6% for controls (51.8 ng/mL to 374.0 ng/mL). Within-Lab (Total Imprecision): CV ranges from 4.0% to 7.2% for samples and 4.5% to 5.5% for controls.
    Linearity/Assay Reportable RangeCLSI EP06-A (linearity of Quantitative Measurement Procedures). Implied: Acceptable percentage difference between observed and predicted values.The linearity data supports an analytical measuring range of 0.9 - 1650 ng/mL. Predicted % Difference (Y-Ŷ)/Ŷ*100: values ranged from -8.88% to 9.47% (excluding the lowest observed value of 0.20 ng/mL which was < LoO).
    Detection Limit (LoB, LoD, LoQ)CLSI EP17-A (Protocols for Determination of Limits of Detection and Limits of Quantitation).LoB: 0.3 ng/mL LoD: 0.7 ng/mL LoQ: 0.9 ng/mL (analyte concentration corresponding to 20% within lab CV).
    Interference (Endogenous & Exogenous)CLSI EP7-A2 guidelines. Implied: No significant interference (< ± 10% bias from control) up to tested concentrations.Endogenous: No significant interference (< ± 10% bias) from Hemoglobin (900 mg/dL), Triglyceride (2000 mg/dL), Conjugated Bilirubin (60 mg/dL), and Un-conjugated Bilirubin (60 mg/dL). Exogenous: No significant interference (< ± 10% bias) from all listed substances (e.g., Heparin, N-acetylcysteine, Acetylsalicylic Acid, Ibuprofen, Prednisone, Ferrous Sulphate, Ascorbic Acid) at tested concentrations.
    Specificity (Cross-reactivity)Not explicitly stated as a numerical acceptance criterion, but implied to be acceptable based on comparison with substances.Liver Ferritin: Cross-Reactivity ranged from 94% to 115% at various ferritin levels (when 285 ng/mL of liver ferritin was added). Spleen Ferritin: Cross-Reactivity ranged from 91% to 103% at various ferritin levels (when 225 ng/mL of spleen ferritin was added).
    Dilution RecoveryImplied: Acceptable recovery range (likely around 90-110%).Recoveries ranged from 89%-100% with a mean of 94% for samples diluted 1:2, 1:4, 1:8, and 1:16.
    Spiking RecoveryImplied: Acceptable recovery range (likely around 90-110%).Recoveries ranged from 90%-116% with a mean of 104% for samples spiked with varying amounts of ferritin.
    High-Dose Hook EffectImplied: No hook effect within the intended assay range or a specified high concentration.The assay did not demonstrate a Hook Effect up to 80,000 ng/mL ferritin.
    TraceabilityTraceable to an international standard.Traceable to World Health Organization 2nd International Standard (WHO 80/578).
    Stability (Reagents)Shelf-life, on-board stability, and calibration frequency must meet sponsor pre-defined acceptance criteria.Shelf-life: 8 months at 2-8°C (unopened). On-board/Open-vial: Reagents stable for 28 days once placed on the system. Calibration frequency: 28 days. Sponsor pre-defined acceptance criteria were met.
    Method Comparison (vs. Predicate)Demonstrates substantial equivalence to the predicate device (ADVIA Centaur Ferritin Assay). Expected good correlation (high r-value) and close to 1 slope with minimal intercept.r = 0.99 (Linear regression) Weighted Deming regression: y = 1.03x - 0.5 ng/mL Sample range: 3.4-1641.4 ng/mL
    Matrix Comparison (Sample Types)Demonstrates equivalence across different sample matrices (serum, K2 EDTA plasma, Lithium Heparin plasma). Expected good correlation (high r-value) and close to 1 slope with minimal intercept.EDTA plasma vs. Serum: y = 0.96x + 1.6 ng/mL, r = 0.997. Lithium heparin vs. Serum: y = 0.95x + 0.1 ng/mL, r = 0.998. Sample range for both: 2.5-1440.7 ng/mL.
    Expected Values/Reference RangeEstablished using appropriate methodology.Established using the Atellica IM Analyzer in accordance with CLSI Document EP28-A3c. Normal Males (N=179): Median 58.9 ng/mL, 95th Percentile Range 10.5-307.3 ng/mL. Normal Females (N=275): Median 46.4 ng/mL, 95th Percentile Range 7.3-270.7 ng/mL.

    Study Details

    1. Sample sizes used for the test set and the data provenance:

      • Precision: 6 serum-based samples, plus calibrators and controls. 80 replicates per sample (20 days, 2 runs/day, duplicate).
      • Linearity/Assay Reportable Range: 11 samples. Tested in triplicates.
      • Detection Limit (LoB, LoD, LoQ):
        • LoB: 6 blank samples, 2 reagent lots, 2 replicates/day over 10 days (System 1) and 9 days (System 2).
        • LoD: 7 low-level human serum samples, 2 reagent lots, 10 days (System 1) and 9 days (System 2).
        • LoQ: 6 low human serum samples (System 1) and 5 low human serum samples (System 2), 2 reagent lots, 8 replicates/day over 5 days.
      • Interference:
        • Endogenous: 2 human serum pools (~20 ng/mL and ~200 ng/mL) spiked with interferent. Tested in replicates of 3.
        • Exogenous: 2 human serum pools (~20 ng/mL and ~200 ng/mL) spiked with interferent. Tested in replicates of 3.
      • Specificity (Cross-reactivity): 4 specimens with different endogenous ferritin levels. Replicates of 3.
      • Dilution Recovery: 3 human serum samples.
      • Spiking Recovery: 5 samples.
      • High-dose hook effect: 11 serial dilution samples from a high-concentration sample.
      • Method Comparison: 126 serum samples (107 included in final calculations).
      • Matrix Comparison: 56 paired sample sets (serum, K2 EDTA plasma, Lithium Heparin plasma).
      • Expected Values/Reference Range: 179 normal males, 275 normal females.

      Data Provenance: The document explicitly states "human serum" and "human serum-based samples". The reference ranges were established using "apparently healthy male and female subjects." The document does not specify the country of origin for the samples and does not explicitly state whether the data was retrospective or prospective, though the nature of laboratory performance studies usually implies prospective collection for the specific tests, but potentially using residual samples for some evaluations.

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

      • This device is an in vitro diagnostic assay for quantitative determination of ferritin. The "ground truth" here is the actual concentration of Ferritin in the samples, measured by a reference method or known by spiking/dilution. It is not typically established by human experts in the same way imaging or pathology devices are. The accuracy of the analytical measurements themselves are the ground truth for these types of studies.
      • For the "Expected Values/Reference Range" study, subjects selected had "normal liver function enzyme tests, bilirubin, and serum iron tests," which would imply medical expert evaluation of their health status, though no specific number or qualification of experts is mentioned for this selection.

    3. Adjudication method for the test set:

      • Not applicable in the context of an in vitro diagnostic assay for quantitative measurement. The results are numerical values from the assay, and "adjudication" in the sense of expert consensus on a diagnosis is not relevant for establishing the assay's analytical performance.

    4. 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:

      • Not applicable. This is an in vitro diagnostic assay, not an AI-powered diagnostic imaging or pathology system that assists human readers.

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

      • Yes, this entire submission describes the standalone performance of the Atellica IM Ferritin Assay (an in vitro diagnostic device). The performance characteristics are measured directly from the assay without human interpretation being part of the primary performance evaluation, beyond operating the instrument and interpreting the numerical output.

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

      • The ground truth for the analytical performance studies (precision, linearity, detection limit, interference, recovery) is either:
        • Known concentrations: Achieved through spiking or dilution with known amounts of ferritin.
        • Reference method comparison: For the method comparison study, the predicate device (ADVIA Centaur Ferritin Assay) serves as the comparative "truth" or reference.
        • Traceability to an international standard: The assay is traceable to the WHO 2nd International Standard (WHO 80/578).

    7. The sample size for the training set:

      • For in vitro diagnostic assays of this type, a "training set" in the machine learning sense is not typically used for establishing the analytical performance characteristics. The assay is based on chemical reactions and optical detection, not a machine learning algorithm that learns from data.
      • However, if you consider the development process, there would have been internal R&D studies using various samples to optimize reagents and calibration, but these are not disclosed as formal "training sets" in the 510(k) submission.

    8. How the ground truth for the training set was established:

      • As explained above, the concept of a "training set" and its associated ground truth in the context of machine learning does not directly apply to the analytical performance studies of this immunoassay. The validity comes from the known chemical properties of the reagents and the calibration against internationally recognized standards.
    Ask a Question

    Ask a specific question about this device

    Page 1 of 1