Search Results

The Access Free T4 assay is a paramagnetic particle, chemiluminescent immunoassay for the quantitative determination of free thyroxine levels in human serum and plasma (heparin) for the diagnosis and treatment of thyroid diseases using the Access Immunoassay Systems.

Assay type: two-step, competitive

The Access Free T4 assay is a two-step enzyme immunoassay. The Access Free T4 assay consists of the reagent pack and calibrators. Other items needed to run the assay include substrate and wash buffer. The Access Free T4 reagent pack, Access Free T4 calibrators, along with Wash Buffer II are designed for use with the Access Immunoassay Systems in a clinical laboratory setting.

The Access Free T4 contains the following components:

• R1a: Dynabeads paramagnetic particles coated with streptavidin and mouse monoclonal anti-Thyroxine (T4) coupled to biotin; preservative
• R1b: TRIS buffered saline with protein (avian), surfactant, preservative
• R1c: TRIS buffered saline with protein (avian), surfactant, preservative.
• R1d: Triiodothyronine-alkaline phosphatase (bovine) conjugate in a TRIS buffer with protein (avian), surfactant, preservative.
• R1e: TRIS buffer with protein (avian and murine), surfactant, preservative

The provided 510(k) summary describes the Access Free T4 assay, a paramagnetic particle, chemiluminescent immunoassay for the quantitative determination of free thyroxine levels in human serum and plasma for the diagnosis and treatment of thyroid diseases.

Here's an analysis of the acceptance criteria and the study that proves the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance

The document outlines acceptance criteria and performance for several studies.

2. Sample Sizes Used for the Test Set and Data Provenance

• Method Comparison: 163 serum samples.
• Imprecision: For both Access 2 and Dxl 9000, "multiple samples" were tested in duplicate in 2 runs per day for a minimum of 20 days. The tables show N values per sample type ranging from 80 to 84 (representing the total number of replicates over the 20+ days for each control sample).
• Detection Capability (LoB, LoD, LoQ): "multiple reagent lots and 3 instruments over a minimum of 3 days" (LoB) and "multiple reagent lots and 3 instruments over a minimum of 5 days" (LoD, LoQ).
• Analytical Specificity: Serum samples with two concentrations of Free T4, tested in replicates of six each.
• Interference: Patient serum samples with two levels of Free T4, with 6 to 12 replicates tested for each control sample preparation.
• Sample Type Comparison: A minimum of 40 matched sets of patient samples were tested with each reagent lot. Specifically, 41 matched sets for Access 2 and 43 matched sets for Dxl 9000 were reported.

Data Provenance: The document does not explicitly state the country of origin of the data or whether it was retrospective or prospective. Given the nature of these clinical performance studies for an in vitro diagnostic device, it is typically prospective, involving controlled testing in a laboratory setting with collected human samples. The term "patient serum samples" suggests human origin.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications

This section is not applicable to this type of device (in vitro diagnostic immunoassay). The "ground truth" for an immunoassay is typically established by reference methods, consensus methods, or highly characterized samples, not by expert human readers. The predicate device itself acts as a comparative standard in the method comparison study.

4. Adjudication Method

This section is not applicable for this type of device as there is no human interpretation or decision-making process that would require adjudication.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

This section is not applicable. This device is an automated immunoassay for quantitative determination of a specific analyte (Free T4), not an imaging or diagnostic device that relies on human reader interpretation (with or without AI assistance).

6. Standalone Performance Study

Yes, the entire set of studies described (Method Comparison, Imprecision, Detection Capability, Linearity, Analytical Specificity, Interference, Sample Type Comparison) represents the standalone performance of the modified device (Access Free T4) when compared against the predicate device or a recognized standard (e.g., CLSI guidelines). The performance metrics provided (slope, correlation, %CV, SD, LoB, LoD, LoQ, % difference) quantify the device's accuracy, precision, and analytical characteristics.

7. Type of Ground Truth Used

The ground truth used depends on the specific study:

• Method Comparison: The predicate device (Access Free T4 Assay on the Access Immunoassay Analyzer K982250) served as the comparator or "reference" for comparison.
• Imprecision: Statistical methods define precision (e.g., repeatability, within-laboratory variability). Control samples with known target values serve as the basis for evaluation.
• Detection Capability (LoB, LoD, LoQ): These limits are derived statistically from measurements of blank samples and low-concentration samples.
• Linearity: Expected proportional changes in analyte concentration across a range are compared against measured values.
• Analytical Specificity and Interference: Known concentrations of potential cross-reactants or interferents added to serum samples with known Free T4 levels are used. The "true" effect is zero interference/cross-reactivity.
• Sample Type Comparison: Matched patient samples (serum vs. plasma) are used, where the expectation is agreement between results from different sample matrices from the same individual.

8. Sample Size for the Training Set

This document only describes analytical performance studies for the modified device. There is no mention of a "training set" in the context of machine learning or AI, as this is an immunoassay device, not a software algorithm that would typically undergo a training phase.

9. How the Ground Truth for the Training Set Was Established

This is not applicable as there is no training set for this type of device.

Ask a specific question about this device

Assay for the in vitro quantitative determination of free thyroxine in human serum and plasma. Measurements obtained by this device are used in the diagnosis and treatment of thyroid disease.

The electrochemiluminescence immunoasay "ECLIA" is intended for use on the cobas e immunoassay analyzers.

The Elecsys FT4 IV immunoassay a fourth generation FT4 assay by Roche Diagnostics for the for the in vitro quantitative determination of free thyroxine in human serum and plasma. It is intended for use on the cobas e immunoassay analyzers. The cobas e family of analyzers uses electrochemiluminescence immunoassay "ECLIA" technology. The assay is an 18 minute assay utilizing a competition principle using a monoclonal antibody which is specifically directed against free thyroxine. Results are determined via a calibration curve which is instrument specifically generated by 2-point calibration against the master curve for that reagent lot.

Here's a breakdown of the acceptance criteria and the study information for the Elecsys FT4 IV device, based on the provided text:

Device: Elecsys FT4 IV (Free Thyroxine Test System)
Predicate Device: Elecsys FT4 II (K131244)
Intended Use: In vitro quantitative determination of free thyroxine in human serum and plasma for the diagnosis and treatment of thyroid disease, intended for use on cobas e immunoassay analyzers.

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state acceptance criteria in a dedicated section with specific numerical targets for each performance characteristic. Instead, it generally states that "All predefined acceptance criteria was met" for various studies. The reported device performance is presented as the results obtained from these studies.

2. Sample Sizes Used for the Test Set and Data Provenance

• Precision (Repeatability and Intermediate Precision): Not explicitly stated how many individual samples were used, but multiple human serum samples (6 levels) and PreciControl Universal (2 levels) were tested. Repeatability involved measurements over 21 days and 5 days, and intermediate precision over the same periods.
• Analytical Sensitivity (LoB, LoD, LoQ):
  • LoB: One blank sample (fT4 depleted human serum pool) with ten replicates per run, across six runs over three or more days, evaluated on three reagent lots. Total 60 determinations.
  • LoD: Five low-level human serum sample pools (diluted) with two replicates/sample/run, across six runs over three or more days, evaluated on three reagent lots.
  • LoQ: At least five low-level samples of serum, with five replicates per sample per run, over five days (total 25 replicates/sample/reagent lot), evaluated on three reagent lots.
• Linearity/Assay Reportable Range: Three individual human serum samples (spiked and diluted) to prepare dilution series. 13 concentrations (levels) prepared. Samples assayed in 4-fold determination within a single run.
• Endogenous Interferences: Nine endogenous substances evaluated. The number of samples for each is not specified, but typically involves spiked samples.
• Analytical Specificity/Cross-Reactivity: For each potential cross-reacting compound, two human serum samples (low and slightly elevated FT4 levels) were tested.
• Exogenous Interferences (Drugs): 17 commonly and 15 specially used pharmaceutical compounds evaluated.
• Sample Matrix Comparison: Samples (native single donors and pools, spiked or diluted) drawn into Serum, Li-Heparin, K2-and K3-EDTA plasma primary tubes. Number of samples not specified.
• Method Comparison to Predicate: 121 serum samples.
• Expected Values/Reference Range: 150 apparently healthy subjects.

Data Provenance:

• For the Expected Values/Reference Range study, serum samples were obtained from a "commercial vendor" and collected from "health donors in the United States."
• Other studies generally refer to "human serum" or "human serum samples," implying samples of human origin, but specific countries or retrospective/prospective nature are not typically detailed for non-clinical analytical performance studies unless relevant regulations require it. However, given the context of an IVD, these are typically laboratory-based studies using banked or ethically sourced human samples.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

The document describes the analytical performance of an in vitro diagnostic (IVD) device. For IVDs measuring biomarkers, "ground truth" is typically established by existing reference methods, certified reference materials, or by the analytical properties of the substance itself (e.g., purified analytes for spiking studies).

• No human experts are explicitly mentioned for establishing ground truth in the context of diagnostic accuracy for the Elecsys FT4 IV performance studies.
• The ground truth for FT4 concentration in samples used for studies like linearity or precision is based on the expected concentration determined by established laboratory practices, dilutions, spiking with known amounts of analyte, or values obtained from a reference method (in the case of method comparison).
• For the Reference Range study, classification as "apparently healthy subjects" serves as a form of ground truth for establishing normal ranges, likely based on medical history or screening criteria provided by the commercial vendor.

4. Adjudication Method for the Test Set

Not applicable. This is an IVD device measuring a quantitative biomarker (free thyroxine). The performance evaluation involves analytical studies (precision, sensitivity, linearity, interference, method comparison) against established analytical standards or a predicate device, not subjective interpretation requiring adjudication by multiple readers or experts.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

No. MRMC studies are typically performed for imaging devices or devices that involve human interpretation of results. This is an IVD that quantitatively measures a biomarker on an automated analyzer. The device produces a numerical result, and there is no "human reader" in the loop for interpreting device output in a way that would necessitate an MRMC study for comparative effectiveness.

6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done

Yes, this is implicitly a standalone study for the device. The entire non-clinical performance evaluation (Sections 4.1 to 4.11) describes the performance of the Elecsys FT4 IV assay itself, as an automated system on the cobas e immunoassay analyzers. It evaluates the device's accuracy, precision, sensitivity, and resistance to interference independently of human interpretation of the final numerical result. The output (a quantitative FT4 value) is what clinicians interpret in the context of a patient's condition.

7. The Type of Ground Truth Used

The ground truth used for these analytical studies primarily consists of:

• Reference Standards/Known Concentrations: For studies like linearity, analytical sensitivity (LoB, LoD, LoQ), and interference, samples are often prepared with known concentrations of the analyte or interfering substances.
• Comparative Reference Method: For the method comparison study, the predicate device (Elecsys FT4 II) served as the reference for comparison.
• Absence of Analyte: For LoB determination, "fT4 depleted human serum sample pool" was used.
• Clinically Defined Status: For the reference range study, samples from "apparently healthy donors" were used to establish normal values, which serves as a clinical ground truth for a non-diseased population.

8. The Sample Size for the Training Set

The document describes an analytical device and its performance verification, not a machine learning or AI algorithm in the traditional sense that requires a "training set" for model development.

• The Elecsys FT4 IV assay is an ECLIA (Electrochemiluminescence Immunoassay), which is a chemical reaction-based detection system, not a software algorithm that learns from data like an AI model.
• Therefore, the concept of a "training set" in the context of AI/ML is not directly applicable here. The device's calibration curve is generated using calibrators ("CalSet FT4 IV") against a master curve for each reagent lot, which is a standard procedure for IVDs, not an AI training process.

9. How the Ground Truth for the Training Set Was Established

As explained in point 8, the concept of a "training set" for an AI/ML model is not applicable to an immunoassay device like the Elecsys FT4 IV.

The "ground truth" for the device's operational parameters (like its calibration curve) is established through:

• Reference materials/calibrators: The device uses specific "CalSet FT4 IV" calibrators. These calibrators have assigned values traceable to reference methods or primary standards, which serve as the "ground truth" for calibrating the device for quantitative measurements.
• Master curve: The instrument-specific calibration curve is generated by 2-point calibration against a master curve for that reagent lot. The master curve itself is derived from extensive characterization of a reagent lot using precisely known reference materials across the measuring range.

Ask a specific question about this device

The MAGLUMI 2000 FT4 assay is for in vitro diagnostic use in the quantitative determination of free thyroxine (FT4) in human serum. The measurement of FT4 is used in the diagnosis of thyroid disorders.

MAGLUMI 2000 FT4 kit consists of the following reagents: Magnetic Microbeads- coated with T4 antigen, BSA, NaN3(

Here's an analysis of the acceptance criteria and study findings for the MAGLUMI 2000 FT4 device, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

Note on Acceptance Criteria: The document often refers to CLSI guidelines (e.g., EP5-A2, EP6-A, EP17-A, EP9-A2) for performance characteristic studies. While explicit numerical acceptance criteria are not always stated directly in the text, compliance with these guidelines implicitly means that the performance falls within industry-accepted ranges and statistical thresholds. For precision, for instance, a total CV of less than 10% (or even lower for higher concentrations) is generally considered acceptable in clinical chemistry. For linearity and method comparison, an R² value close to 1, a slope near 1, and a y-intercept near 0 are standard indicators of good performance.

2. Sample Size Used for the Test Set and Data Provenance

• Precision Study:
  • Sample Size: 240 measurements per control/calibrator/pool (N=240, from 80 samples analyzed per level on each of 3 instruments).
  • Data Provenance: Not explicitly stated, but implied to be patient serum pools and native patient samples from clinical settings, likely domestic to the manufacturer's testing location or within industry standard practices for clinical evaluation. The study was conducted on "three controls, two calibrators, four spiked patient serum pools and four native patient sample pools."
• Linearity Study:
  • Sample Size: 11 levels with FT4 concentrations from 0.19 to 10.0 ng/dL, each measured in quadruplicate on 3 lots of reagent.
  • Data Provenance: Samples prepared by spiking T4 USP Standard into T4-free human serum samples.
• Detection Limit Study (LOB, LOD, LOQ):
  • LOB: 80 measurements of T4 free human serum samples.
  • LOD: Four levels of low samples, measured in 80 replicates over 5 days per sample.
  • LOQ: Six low serum samples, in six replicates per run, one run per day, over 5 days.
  • Data Provenance: T4 free human serum samples and low serum samples.
• Interference Study:
  • Sample Size: Variable, involves preparing solutions, and for common drugs/interfering substances, three serum samples (low, medium, high FT4) per substance. For HAMA/RF/Protein, specific FT4 human serum samples.
  • Data Provenance: Spiked solutions, human serum pools, FT4 human serum samples.
• Method Comparison Study:
  • Sample Size: 224 human serum samples.
  • Data Provenance: Human serum samples. Provenance (e.g., country of origin, retrospective/prospective) is not explicitly stated.
• Expected Values/Reference Range Study:
  • Sample Size: 157 serum samples.
  • Data Provenance: Normal, apparently healthy adult individuals (22 years and older). Provenance (e.g., country of origin, retrospective/prospective) is not explicitly stated.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

This information is not provided in the document. For an in vitro diagnostic device like the MAGLUMI 2000 FT4, the "ground truth" for analytical performance tests (precision, linearity, detection limits, interference) is typically established by the reference methods, spiked concentrations, or consensus values derived from the testing procedures themselves, rather than human expert opinion. For method comparison, the "ground truth" is typically the result from the established predicate device.

4. Adjudication Method for the Test Set

The concept of an "adjudication method" (like 2+1 or 3+1 for clinical diagnoses based on expert review) is not applicable to the analytical performance studies described for this in vitro diagnostic device. The studies rely on quantitative measurements and statistical analysis rather than subjective expert interpretation of results for ground truth.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done

No, an MRMC comparative effectiveness study was not done. This type of study is primarily relevant for medical imaging devices or other diagnostic tools where human readers interpret results, and the AI system acts as an aid to improve human performance. For an automated in vitro diagnostic assay like MAGLUMI 2000 FT4, the performance is evaluated through direct analytical comparisons.

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

Yes, the studies described are for the standalone performance of the MAGLUMI 2000 FT4 device. These are analytical performance studies (precision, linearity, detection limits, interference, method comparison) that evaluate the device's ability to accurately and precisely measure FT4 concentrations in human serum, independent of human interpretation or intervention in the measurement process itself. The "algorithm" here refers to the immunoassay's chemistries and the instrument's measurement principles.

7. The Type of Ground Truth Used

• Precision: Reference materials (controls, calibrators) with established target values, and pooled human serum samples where the mean serves as a statistical "ground truth" for variability assessment.
• Linearity: Gravimetrically or volumetrically prepared "expected" concentrations of T4 in serum.
• Stability: Initial measurements at time 0, with subsequent measurements compared to these baselines.
• Detection Limits (LOB, LOD, LOQ): Analyte-free samples and low-concentration samples measured extensively to statistically determine minimum detectable/quantifiable levels.
• Interference (Cross-Reactivity): Known amounts of potential interferents/cross-reactants.
• Method Comparison: Results from the predicate device (Siemens ADVIA Centaur FT4) are used as the reference "ground truth" for comparison.
• Expected Values/Reference Range: Statistical distribution of measurements from a reference population (157 apparently healthy adults) to establish a normal range.

8. The Sample Size for the Training Set

This information is not provided. The document describes performance characteristic studies (test set), but does not detail the development or "training" specific to an AI algorithm for this device. For in vitro diagnostics, "training set" would typically refer to samples used during the assay development and optimization phase to establish parameters like calibration curves, reagent formulations, or instrument settings. The document focuses on validation studies.

9. How the Ground Truth for the Training Set Was Established

This information is not provided as a "training set" isn't explicitly defined in the context of this submission. If referring to the development/optimization phase of a typical immunoassay, ground truth would be established through:

• Reference materials: Highly purified FT4 standards.
• Known concentrations: Samples prepared with precise, known concentrations of FT4.
• Comparative data: Initial comparisons against existing reference methods or well-characterized assays during the research and development phase.
• Clinical samples: Use of a large number of clinical samples to optimize the assay's dynamic range and specificity.

Ask a specific question about this device

Assay for the in vitro quantitative determination of free thyroxine in human serum and plasma. Measurements obtained by this device are used in the diagnosis and treatment of thyroid disease.

The electrochemiluminescence immunoassay "ECLIA" is intended for use on the cobas e 411 immunoassay analyzer.

The Elecsys FT4 III immunoassay is used for the in vitro quantitative determination of free thyroxine in human serum and plasma. It is intended for use on the cobas e 411 immunoassay analyzer. The cobas e family of analyzers uses electrochemiluminescence immunoassay "ECLIA" technology. The reagent working solutions include: Rackpack (kit placed on analyzer) M: Streptavidin-coated microparticles. R1: Anti-T4-Abbiotin. R2: Anti-T4-AbRu(bpy).

This document describes the premarket notification for the Elecsys FT4 III device, an immunoassay for the quantitative determination of free thyroxine. The submission details the non-clinical performance evaluation of the device as it relates to its substantial equivalence to a predicate device.

Here's an analysis of the provided information regarding acceptance criteria and supporting studies:

1. Table of Acceptance Criteria and Reported Device Performance

The document provides a comparative assay table between the Elecsys FT4 II (predicate device) and Elecsys FT4 III (candidate device), and also presents detailed studies for various performance characteristics. I will synthesize the acceptance criteria (often implied by the "same" or within a stated range compared to the predicate, or by specific criteria in the study descriptions) and the reported performance as presented in the document.

2. Sample Size Used for the Test Set and the Data Provenance

• Precision (Repeatability and Intermediate Precision):

  • Sample Size: Two levels of control (PreciControl Universal) and five human sera (native single donors, native serum pools, spiked serum pools). Tested in 2 replicates per run, 2 runs per day for 21 days.
  • Data Provenance: Not explicitly stated, but "human sera" and "single donors" suggest human samples, likely retrospective pools or prospectively collected for the study. No country of origin mentioned.

• Endogenous Interferences (Biotin):

  • Sample Size: Three human serum samples containing low, mid, and high concentrations of fT4, tested in accordance with CLSI EP07-A2.
  • Data Provenance: Human serum samples (single donors, native as well as spiked).

• Sample Matrix Comparison:

  • Sample Size: 53 serum/plasma pairs per sample material (serum, K2-EDTA plasma, K3-EDTA plasma).
  • Data Provenance: Native single donors and pools, as well as spiked samples; human.

• Method Comparison to Predicate:

  • Sample Size: 141 serum samples (138 native human serum samples and 3 spiked sample pools).
  • Data Provenance: Human serum samples.

• Reagent Stability (After First Opening, On-board, Alternate Storage):

  • Sample Size: Six human serum (HS) samples (native single donor serum, native serum pools, and spiked serum pools) and two controls (PreciControl Universal). Tested in duplicate (for "first opening" and "alternate storage") or two-fold determination (for "on-board").
  • Data Provenance: Human serum samples.

• Reagent Real-time Stability:

  • Sample Size: Three human serum samples.
  • Data Provenance: Human serum samples (stored at -20°C).

• Calibration Stability (Lot and On-board):

  • Sample Size: Five human serum (HS) samples (single donor serum, native serum pools, and spiked serum pools) and two controls (PreciControl Universal). Tested in two-fold determination.
  • Data Provenance: Human serum samples.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts

This document describes the performance evaluation of an in vitro diagnostic (IVD) device, specifically an immunoassay for free thyroxine. For such devices, "ground truth" is typically established by reference methods, consensus values for control materials, or by testing against well-characterized clinical samples in comparison to an established predicate device. It is not generally established by a panel of human experts reviewing individual cases in the same way an imaging AI algorithm might be.

Therefore, the concept of "number of experts used to establish the ground truth" as it applies to imaging or clinical decision support AI is not applicable here. The ground truth for these types of studies is based on biochemical measurements and analytical comparisons.

4. Adjudication Method for the Test Set

Again, for an IVD device evaluating a biomarker, traditional "adjudication" by human experts in the context of clinical interpretation of individual cases (e.g., 2+1, 3+1 for discordances in imaging) is not applicable. The performance is adjudicated against analytical measurements and statistical criteria defined in CLSI guidelines and internal validation plans.

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

Not Applicable. This is an in vitro diagnostic device (immunoassay), not an AI-powered clinical decision support or imaging interpretation tool that "assists human readers." Therefore, no MRMC study with human readers improving with AI assistance would be performed.

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

The Elecsys FT4 III is an automated immunoassay intended for use on the cobas e 411 analyzer. Its performance is inherently "standalone" in the sense that the analyzer and reagents perform the measurement autonomously. A human laboratory professional loads samples and reagents, and reviews the results. The studies described (precision, interference, method comparison) are all evaluating the standalone analytical performance of this device.

Therefore, yes, the studies described inherently evaluate the "standalone" performance of the algorithm/device system in quantifying free thyroxine in samples.

7. The Type of Ground Truth Used

The ground truth for the various studies is established through:

• Reference methods / Predicate Device: For method comparison, the Elecsys FT4 II assay (K131244) served as the reference method ("X").
• Known concentrations: For interference studies (e.g., biotin, cross-reactivity), samples were spiked with known concentrations of interfering substances or analytes.
• Industry Standards / Guidelines: Adherence to CLSI guidelines (e.g., EP05-A3 for precision, EP07-A2 for interferences) implies that the ground truth for acceptability criteria is based on established laboratory practice and industry consensus for robust analytical performance.
• Controls and Calibrators: PreciControl Universal and CalSet FT4 III are used, which have established target values and ranges for quality control and calibration, respectively.

8. The Sample Size for the Training Set

This document describes the performance validation of a commercially manufactured in vitro diagnostic device, not the development of a machine learning algorithm. Therefore, the concept of a "training set" in the context of AI model development is not directly applicable here.

The device's underlying principles are based on established electrochemiluminescence immunoassay (ECLIA) technology and biochemically derived reagent formulations. While there is a development process that optimizes reagents and assay parameters, this is not typically referred to as "training" in the AI sense, nor does the document provide details about such a "training set" if it were analogous.

9. How the Ground Truth for the Training Set Was Established

Given that "training set" as understood in AI/ML model development is not directly applicable to this IVD device's validation data, this question is not applicable in the context of the provided document. The "ground truth" for the device's development (the underlying chemical and physical principles on which it operates) would have been established through extensive biochemical research, assay development, and optimization processes, rather than a discrete "training set" with established ground truth labels in the AI sense.

Ask a specific question about this device

The ARCHITECT Free T4 (FT4) assay is a Chemiluminescent Microparticle Immunoassay (CMIA) for the quantitative determination of free thyroxine (Free T4) in human serum and plasma.

ARCHITECT Free T4 (FT4) assay is to be used as an aid in the assessment of thyroid status.

The ARCHITECT Free T4 assay is a two-step immunoassay for the quantitative determination of free thyroxine (Free T4) in human serum and plasma using CMIA technology with flexible assay protocols, referred to as Chemiflex.

In the first step, sample and anti-T4 coated paramagnetic microparticles are combined. Free T4 (unbound) present in the sample binds to the anti-T4 coated microparticles. After washing, Ts acridinium-labeled conjugate is added to create a reaction mixture. Following another wash cycle, pre-trigger and trigger solutions are added to the reaction mixture. The resulting chemiluminescent reaction is measured as relative light units (RLUs). An inverse relationship exists between the amount of Free T4 in the sample and the RLUs detected by the ARCHITECT i System optics.

The provided text describes a 510(k) premarket notification for a modification to the ARCHITECT Free T4 assay, not a study proving device meets acceptance criteria in the context of AI/ML or comparative effectiveness. The device is an in-vitro diagnostic (IVD) test, and the information requested about AI/ML models, human readers, and ground truth types (pathology, outcomes data) is not applicable to this type of device or documentation.

However, I can extract the acceptance criteria and performance data related to the nonclinical performance verification for the modified ARCHITECT Free T4 assay, as presented in the document.

Acceptance Criteria and Reported Device Performance

The device under review is an in-vitro diagnostic (IVD) assay, not an AI/ML device. Therefore, acceptance criteria and performance are based on analytical performance studies. The document states that the "device passed all of the tests based on pre-determined acceptance criteria." While the specific numerical acceptance criteria are not detailed in this summary, the types of tests performed and the general outcome are reported.

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

The document does not specify the exact sample sizes used for each individual nonclinical performance study (Limit of Blank/Detection/Quantitation, Precision, Accuracy, Stability, Linearity). It also does not explicitly state the country of origin of the samples or if they were retrospective or prospective, though for IVD analytical performance, these would typically be control samples and patient samples collected for various analytical evaluations to represent the intended use population.

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

Not applicable. For an immunoassay, the "ground truth" for analytical performance studies is established by the assay's ability to accurately and precisely measure the analyte (Free T4) in samples against known concentrations or reference methods, not by expert interpretation of images or clinical data.

3. Adjudication method (e.g. 2+1, 3+1, none) for the test set:

Not applicable. Adjudication methods like 2+1 are used for expert consensus on subjective interpretations (e.g., radiology reads), which is not relevant for an IVD analytical performance study.

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 IVD assay, not an AI/ML diagnostic tool involving human readers.

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

Not applicable. This is an IVD immunoassay, not an algorithm or AI model. Its performance is inherent to the assay chemistry and instrument.

6. The type of ground truth used:

For the analytical performance studies, the "ground truth" is established through:

• Known concentrations: For studies like linearity, limits, and precision, samples with known concentrations of Free T4 would be used.
• Reference methods/comparative methods: For accuracy by correlation, the results of the ARCHITECT Free T4 assay would be compared against a legally marketed predicate device (as referenced, K123379) or a recognized reference method.

7. The sample size for the training set:

Not applicable. This is an immunoassay, not an AI/ML model, so there is no "training set" in the computational sense. The development of the assay reagents and protocols is based on chemical and biological research and optimization.

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

Not applicable, as there is no "training set" in the context of an AI/ML model. The development of the assay components (reagents, microparticles, conjugate) and the associated manufacturing process enhancement (reduced microparticle percent solids) are validated through the nonclinical performance studies listed.

Ask a specific question about this device

The FREND™ Free T4 Test System is a rapid indirect competitive immunoassay for the quantitation of free thyroxine (FT4) in human serum and lithium heparinized plasma specimens using the FREND™ Free T4 system. Measurements of free thyroxine (FT4) are used in the diagnosis of thyroid disorders. The FREND™ Free T4 Test System is intended for use in clinical laboratories. For in vitro diagnostic use only. The test is not intended for point-of-care facilities.

The FREND™ Free T4 is a rapid fluorescence immunoassay that measures Free T4 in human serum and in lithium heparinized plasma using the FREND™ Free T4 Test System. The FREND™ Free T4 is a single use fluorescence immunoassay designed to quantify the concentration of free thyroxine in serum and lithium heparinized plasma samples, The FREND™ Free T4 test is a two-step competitive immunoassay with gold nanoparticles labeled with T4-specific monoclonal anti-T4- antibody (mouse), T3-BSA labeled with fluorescent nanoparticles, and fluorescence detection by the FREND™ System. The FREND™ Free T4 Test utilizes microfluidic technology and detects immunecomplexes bound to Free T4. A 70ul Sample is first incubated during Step 1 for five minutes at 37 degrees C in the Free T4 Gold AB Tube with monoclonal anti-T4 antibody conjugated with gold nanoparticles. In Step 2, 35 ul of the mixture from Step 1 is manually loaded into the inlet of the cartridge, where it hydrates a T3-BSA fluorescent bead conjugate and migrates along the test strip. During migration the bound Free T4 in the sample and the fluorescent bead conjugates of T3-BSA compete to form antigenantibody complex in the test zone. Unbound T3-BSA fluorescent conjugates flow through and bind to the anti-T4 antibody that is fixed on the surface in the reference zone. Step 2 takes approximately four minutes after which the fluorescent signals in the test and reference zones are measured. Free T4 quantification is based upon the ratio of the intensity of the test and reference zones. A lower ratio of fluorescence is indicative of a higher Free T4 concentration, in other words, the maqnitude of the fluorescence ratio is inversely proportional to the amount of Free T4 in the sample. The free T4 detection range of the FREND™ Free T4 Test System is 0.4 to 6.0 ng/dL. Results are determined via a lot-specific calibration curve which is generated by the manufacturer using a six-point calibration determined from values averaged from five replicates at each level. The established curve is uploaded to the FREND™ System via the Free T4 Code-chip and is valid until the lot expiration date. The established curve is saved in the code-chip and valid until the expiration date of the test cartridge lot. The FREND™ Free T4 Test cartridge is a disposable plastic device that houses the reagents and contains a port or opening (inlet) where the sample is applied. Once the sample is applied, it will mix with the reagents and travel towards the detection area via capillary action. The FREND™ System is a portable, automated FREND™ cartridge reader. The FREND™ System is based on quantitative immunoassay technology capable of quantifying single or multiple analytes by measuring laser-induced fluorescence in a single-use disposable reagent cartridge. The FREND™ cartridge utilizes micro-fluidics lateral flow technology where the analyte of interest in the sample forms immune complexes while moving through the fluidics pathway in the cartridge. The concentration of the analyte of interest in an unknown sample is calculated using the ratio of the fluorescent intensity of the test zone and the reference zone. The FREND™ System is a bench top fluorescence reader containing a touch screen user interface. The System has a slot that accepts the FREND™ Free T4 Test Cartridge (which contains the reagents and sample), and is programmed to analyze the Test when the sample has fully reacted with the on-board in-cartridge reagents. Results of the test are displayed on the screen and can be printed on an optional printer. The FREND™ System software controls the graphical user interface, communication with hardware, database manaqement and data analysis. The software also controls the functions of the mechanical components including the motor, laser, printer control and acquisition of data from the sensor. The user can set the time and date and enter patient ID through the graphic user interface. The user cannot make any changes to the software.

Here's a summary of the acceptance criteria and study details for the FREND™ Free T4 Test System, based on the provided 510(k) submission:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are not explicitly stated as pass/fail values in many sections, but rather implied by the statistical analyses meeting CLSI guidelines and demonstrating comparability to the predicate device. The performance data is presented against these implicit standards.

2. Sample Size Used for the Test Set and Data Provenance

• Precision Study: 3 serum pools
• Linearity Study: 1 serum base pool diluted to 11 levels
• Interference Studies: Two levels of free T4 for various endogenous substances and pharmaceuticals.
• Cross-Reactivity Study: Two concentrations for various cross-reactants.
• Method Comparison: 358 samples
• Matrix Comparison: 48 paired serum and lithium heparin samples
• Expected Values/Reference Range: 196 normal, apparently healthy adult individuals

The specific country of origin for the test set data is not explicitly stated, but the precision, linearity, interference, and matrix comparison studies were performed at the NanoEnTek laboratory. The method comparison study was performed at a CLIA-certified laboratory testing facility. The clinical samples for the method comparison were evaluated at that CLIA laboratory. The reference range study used samples from 196 normal, apparently healthy adult individuals. The nature of "clinical samples" for the method comparison implies they would be patient samples, and "normal, apparently healthy adult individuals" for the reference range are also patient-like samples. The studies appear to be prospective in nature, as they involve testing the device with specific protocols (e.g., CLSI guidelines).

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

• None stated for the analytical performance studies. The ground truth for analytical studies (precision, linearity, LoD, LoQ, interference, cross-reactivity) is generally based on the inherent properties of the samples and the performance of established reference methods or spiked concentrations.
• For Method Comparison: The ground truth was established by a predicate device, the Abbott ARCHITECT Free T4 Assay. The number of human experts involved in interpreting results from either the predicate device or the FREND™ Free T4 Test System is not specified, as this is a quantitative immunoassay where results are read by a machine.

4. Adjudication Method for the Test Set

Not applicable for this type of quantitative immunoassay performance study. Ground truth is established by reference methods or defined concentrations, not human adjudication.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and Effect Size

No, a multi-reader multi-case (MRMC) comparative effectiveness study was not done. This device is a quantitative immunoassay, not an imaging device requiring human interpretation, so MRMC studies are not relevant.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Yes, the studies described are standalone performance studies for the FREND™ Free T4 Test System. The device is intended for use in clinical laboratories, and the results are quantitatively measured by the FREND™ System directly. There is no human-in-the-loop performance evaluation described beyond standard laboratory procedures for operating the instrument and handling samples.

7. The Type of Ground Truth Used

• Analytical Performance:
  • Precision, Linearity, LoD, LoQ, Interference, Cross-reactivity: Ground truth was established by prepared samples with known concentrations, spiking experiments, and measurements against established laboratory protocols and standards (CLSI guidelines).
• Method Comparison: Ground truth was established by the predicate device, the Abbott ARCHITECT Free T4 Assay (K123379). This is a comparison against an existing, legally marketed device.
• Matrix Comparison: Ground truth (or comparative standard) derived from measuring the same analyte in different matrices (serum vs. plasma), with the expectation of comparable results.

8. The Sample Size for the Training Set

The document does not explicitly mention a "training set" in the context of an AI/machine learning algorithm. For this immunoassay, the system's "training" involves the manufacturer-generated lot-specific calibration curve (six-point calibration determined from values averaged from five replicates at each level), which is then uploaded to the FREND™ System via a Code-chip.

9. How the Ground Truth for the Training Set Was Established

As described above, the "ground truth" for the device's operational curve (its calibration) is established by:

• Internal standards prepared according to CLSI C45-A Measurement of Free Thyroid Hormones.
• Gravimetric methods based on L-Thyroxine (Sigma T1775, Cell culture grade).
• Confirmation of calibrator Free T4 levels by measurement on the ARCHITECT i free T4 assay (K123379).

This process ensures that the device's internal calibration accurately reflects known Free T4 concentrations.

Ask a specific question about this device

VIDAS® FT4 is an automated quantitative enzyme immunoassay for use on the instruments of the VIDAS® family, for the determination of free thyroxine (FT4) in human serum or plasma (lithium heparin), using the ELFA technique (Enzyme Linked Fluorescent Assay). Measurement of Free Thyroxin is intended for use as an aid in the diagnosis and treatment monitoring of thyroid disorders.

VIDAS® FT4 is an automated quantitative enzyme immunoassay for use on the instruments of the VIDAS® family, for the determination of free thyroxine (FT4) in human serum or plasma (lithium heparin), using the ELFA technique (Enzyme Linked Fluorescent Assay). Measurement of free thyroxine is intended for use as an aid in the diagnosis and treatment monitoring of thyroid disorders.

The assay principle combines a one-step enzyme immunoassay competition method with a final fluorescent detection (ELFA).

The Solid Phase Receptacle (SPR®) serves as the solid phase as well as the pipetting device for the assay. Reagents for the assay are ready-to-use and pre-dispensed in the sealed reagent strips. All of the assay steps are performed automatically by the instrument. The reaction medium is cvcled in and out of the SPR several times.

The sample is collected and transferred into the well containing an alkaline phosphataselabeled anti-T4 antibody (conjugate). The antiqen present in the sample and the T4 antigen coated on the interior of the SPR compete for the available sites on the specific anti-T4 antibody conjugated to alkaline phosphatase.

During the final detection step, the substrate (4-Methyl-umbellifery) phosphate) is cvcled in and out of the SPR. The conjugate enzyme catalyzes the hydrolysis of this substrate into a fluorescent product (4-Methyl-umbelliferone) the fluorescence of which is measured at 450 nm. The intensity of the fluorescence is inversely proportional to the concentration of free thyroxine present in the sample. At the end of the assay, results are automatically calculated by the instrument in relation to the calibration curve stored in memory, and then printed out.

The provided text describes the performance of the VIDAS® FT4 device, an automated quantitative enzyme immunoassay for free thyroxine. However, it does not explicitly state "acceptance criteria" as a separate, pre-defined set of thresholds. Instead, it presents various performance study results (e.g., precision, linearity, correlation, analytical sensitivity, specificity) which implicitly demonstrate that the device meets the necessary performance expectations for its intended use and for demonstrating substantial equivalence to its predicate device.

1. Table of Acceptance Criteria and Reported Device Performance

As explicit acceptance criteria are not tabulated in the document, I will infer the performance metrics considered crucial from the studies presented. For "Acceptance Criteria," I will describe the objective assessed in each study.

2. Sample Sizes Used for the Test Set and Data Provenance

Due to the nature of an in vitro diagnostic (IVD) device, the "test set" is not a singular set of patient cases for a diagnostic decision, but rather refers to various samples used across different analytical and clinical performance studies. The data provenance is generally not explicitly stated as "country of origin" for these types of analytical validation studies but implies laboratory testing within the manufacturer's R&D facilities or contracted labs, primarily related to the device itself.

• Analytical Specificity (Interferences): For common interferences (bilirubin, lipids, hemoglobin, HAMA, albumin), the study involved serum samples at FT4 analyte levels close to the lower and higher limits of the euthyroid range. The number of samples for each interference type is not specified beyond "samples." For drugs, the number of samples is not specified. For structurally related molecules, the study involved tested FT4 concentrations of approximately 1.2 ng/dL and 2.5 ng/dL, but the number of samples is not explicitly given. This data is likely generated prospectively in a laboratory setting.
• Analytical Sensitivity (LoB, LoD, LoQ):
  • LoB: One blank sample, N=60 measures per lot on each instrument (2 lots on one VIDAS®, 1 lot on one miniVIDAS®).
  • LoD: Low concentration samples (number not specified for VIDAS®, 4 samples for miniVIDAS®), N=150 measures per lot on VIDAS®, N=100 measures per lot on miniVIDAS®.
  • LoQ: 9 low concentration samples, N=40 measures per sample and per lot on each instrument (2 lots on one VIDAS®, 1 lot on one miniVIDAS®).
  • This data is prospectively generated in a laboratory setting.
• Linearity: The study was conducted on the VIDAS® and miniVIDAS® instruments. The number of samples or points tested is not explicitly stated. This data is prospectively generated in a laboratory setting.
• Matrix Comparison: 31 sample sets (each set from one donor during one draw, consisting of a reference tube type and four different blood collection tubes). Samples ranged from 0.13 ng/dL to 6.45 ng/dL. This data is prospectively generated.
• Precision: Panel members covering the measuring range were tested. 6 samples for VIDAS® instruments, and 5 samples for miniVIDAS® instruments. Total N=240 measures for each sample/instrument type. This data is prospectively generated in a laboratory setting.
• Correlation: 54 samples were included in the study for comparison to a commercially available Free T4 EIA. Range tested: VIDAS® (0.41 - 6.42 ng/dL), miniVIDAS® (0.52 - 6.35 ng/dL). This data is prospectively generated.
• Reference Range / Expected Values: 544 apparently healthy subjects. The population characteristics are described (45.5% males, 55.5% female, 83.8% Caucasian, 6.1% African-American, 9.6% Hispanic and 0.6% Asian). The country of origin is not specified but the demographic breakdown suggests a Western population. This is prospective data collection.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

For an IVD device like VIDAS® FT4, "ground truth" is typically established by:

• Using reference materials with assigned values.
• Comparing to a legally marketed predicate device or a well-established, validated method (e.g., mass spectrometry, if applicable, though not explicitly mentioned here as a primary reference method).
• For clinical studies like reference range determination, it relies on clinical assessment of health status (e.g., "apparently healthy subjects").

In this documentation:

• Metrological traceability: The VIDAS® FT4 assay is standardized against the Elecsys® FT4 assay (Roche Diagnostic), which serves as a clinical reference method (predicate device).
• Value assignment procedure for calibrator (S1) and control (C1): Reference Calibrators are assigned values via a method comparison between VIDAS® FT4 and Roche Elecsys FT4.
• Correlation study: The VIDAS® FT4 assay was compared to a commercially available Free T4 EIA (likely the predicate Elecsys® FT4 Assay or similar). This serves as the ground truth for that specific study.
• Reference Range / Expected Values: This ground truth is based on the clinical status of "apparently healthy subjects." There is no mention of "experts" in the sense of clinicians or radiologists establishing ground truth for individual cases. The health status of the subjects would be determined by standard medical screening procedures, presumably overseen by medical professionals (e.g., physicians, lab personnel) but not explicitly detailed as "experts."

Therefore, there isn't a stated number of experts or specific qualifications in the traditional sense of consensus reading for image or pathology-based diagnostics. The ground truth relies on established analytical methods and clinical definitions of health.

4. Adjudication Method for the Test Set

Not applicable in the context of this IVD device's analytical and clinical performance studies as described. Adjudication (e.g., 2+1, 3+1) is typically used for resolving discrepancies in expert interpretations in studies involving subjective assessments (like radiology or pathology image analysis). For quantitative assays, the "ground truth" is determined by reference methods or validated values, not by expert consensus on individual results.

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

Not applicable. This is an in vitro diagnostic (IVD) device for quantitative measurement of free thyroxine (FT4). It is not an AI-assisted diagnostic imaging or pathology device that involves human readers interpreting cases.

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

Yes, the studies presented are all standalone performance evaluations of the VIDAS® FT4 assay system (instrument + reagents + algorithm), which is an automated system designed to provide quantitative results directly. The results are "automatically calculated by the instrument in relation to the calibration curve stored in memory, and then printed out." Human involvement is limited to operating the instrument, performing calibration, and reviewing the generated reports, not to interpreting raw data points or making subjective diagnostic decisions the device is intended to automate.

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

The ground truth used primarily consists of:

• Reference Method Comparison: For several studies, the predicate Elecsys® FT4 assay or another commercially available, validated Free T4 EIA served as the comparative "ground truth" or reference method for demonstrating substantial equivalence.
• Reference Materials: For analytical sensitivity (LoB, LoD, LoQ) and linearity, the ground truth is based on gravimetrically prepared or otherwise certified reference materials with known concentrations.
• Clinical Definition: For the reference range determination, the ground truth for the healthy population was based on subjects deemed "apparently healthy."

8. The sample size for the training set

This document describes the validation of an IVD assay, not a machine learning or AI algorithm in the context of "training data." The assay relies on a pre-defined chemical reaction and detection method, with instrument calibration. Therefore, there is no "training set" in the sense of data used to train a predictive model. The "calibration curve" is established using a set of calibrators with known concentrations, which is a standard procedure for quantitative assays and not an AI training process.

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

As there is no "training set" for an AI algorithm in this context, this question is not applicable. For the calibration of the assay, the "ground truth" (assigned values for calibrators) is established through reference method comparison with the Elecsys® FT4 assay and potentially other established metrological traceability chains, as described under "Value assignment procedure for calibrator (S1) and control (C1)."

Ask a specific question about this device

The FT4L method is an in vitro diagnostic test for the quantitative measurement of Free Thyroxine in human serum and plasma on the Dimension® EXL™ integrated chemistry system with LOCI® Module. Measurements of free thyroxine are used in the diagnosis and monitoring of thyroid disease.
The TSHL method is an in vitro diagnostic test for the quantitative measurement of Thyroid Stimulating Hormone (TSH, thyrotropin) in human serum and plasma on the Dimension® EXL™ integrated chemistry system with LOCI® Module. Measurements of TSH are used in the diagnosis and monitoring of thyroid disease.

The Dimension® LOCI Free Thyroxine Flex® reagent cartridge, FT4L and Dimension® LOCI Thyroid Stimulating Hormone Flex® reagent cartridge, TSHL are in vitro diagnostic tests for the quantitative measurement of Free Thyroxine and Thyroid Stimulating Hormone, respectively, in human serum and plasma on the Dimension® EXL™ integrated chemistry system with LOCI® Module. The submission is for the inclusion of pediatric reference intervals to the labeling (Package Inserts) of these assays. No changes were made to the reagents, device design, or manufacturing process.

The provided document is a 510(k) summary for the Siemens Healthcare Diagnostics Dimension® LOCI Free Thyroxine Flex® Reagent Cartridge (FT4L) and Dimension® LOCI Thyroid Stimulating Hormone Flex® Reagent Cartridge (TSHL). The focus of this submission is the addition of pediatric reference intervals to the existing devices, not a primary submission for a new device. Therefore, the information provided primarily pertains to the establishment of these pediatric reference intervals.

Here's an analysis based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria for the pediatric reference intervals are defined by the statistical methods used to establish them and the resulting intervals. The performance reported is the calculated pediatric reference intervals.

Ask a specific question about this device

The FT4L method is an in vitro diagnostic test for the quantitative measurement of Free Thyroxine in human serum and plasma on the Dimension® EXL™ integrated chemistry system with LOCI® Module. Measurements of free thyroxine are used in the diagnosis and monitoring of thyroid disease.

The TSH method is an in vitro diagnostic test for the quantitative measurement of Thyroid Stimulating Hormone in human serum and plasma on the Dimension Vista® System. Measurements of thyroid stimulating hormone produced by the anterior pituitary are used in the diagnosis of thyroid or pituitary disorders.

The submission describes the Dimension® Vista Free Thyroxine Flex® reagent cartridge, FT4 and the Dimension Vista® Thyroid Stimulating Hormone Flex® reagent cartridge, TSH. These are in vitro diagnostic tests for the quantitative measurement of Free Thyroxine and Thyroid Stimulating Hormone in human serum and plasma, respectively, on the Dimension Vista® System. The submission is for the inclusion of pediatric reference intervals to the labeling (Package Inserts) of these assays. No changes were made to the reagents, device design, or manufacturing process.

The document describes the establishment of pediatric reference intervals for the Dimension Vista® Free Thyroxine (FT4) and Thyroid Stimulating Hormone (TSH) Flex® reagent cartridges. It does not contain an acceptance criteria table related to the device's performance against specific metrics of accuracy or precision in the traditional sense, but rather focuses on establishing biological reference intervals. The study's primary goal was to define these pediatric reference intervals.

Here's an analysis of the provided information:

1. Table of Acceptance Criteria and Reported Device Performance

As noted, the document doesn't provide a table of discrete acceptance criteria for device performance (e.g., sensitivity, specificity, accuracy against a gold standard). Instead, the acceptance in this context relates to the establishment and statistical validation of reference intervals for pediatric populations.

The reported "performance" for the Dimension Vista® FT4 and TSH assays, in the context of this submission, is the established pediatric reference intervals. These are compared against adult reference intervals and the assay's analytical measuring range to demonstrate suitability for pediatric use.

Justification of Suitability for Pediatric Use (Acceptance Criteria Implicitly Met):

The document states two key rationales for accepting the pediatric reference intervals, implicitly acting as "acceptance criteria" for the device's application in this population:

• 1. Alignment with Indications for Use: "Testing of pediatric patients is within the established indications for use in the diagnosis and treatment of thyroid disease..."
• 2. Analytical Compatibility: "The newly-established pediatric reference intervals are either within or are above the previously-established reference intervals for euthyroid (normal thyroid) adult populations and they are within the analytical measuring ranges of the Dimension Vista® FT4 and TSH assays. Therefore, the Dimension Vista® FT4 and TSH assays have appropriate analytical performance to test pediatric patients."

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: A total of 421 patients were analyzed to establish the pediatric reference intervals:
  • Infants (01 - 23 months): 82
  • Children (02 - 12 years): 191
  • Adolescents (13 – 20 years): 148
• Data Provenance: The document does not explicitly state the country of origin or if the data was retrospective or prospective. It implies the samples were collected specifically for this study ("Data from a total of 421 patients... were analyzed to establish..."). Given the context of establishing reference intervals, it is likely prospective collection from healthy individuals within the specified age groups, though this is not explicitly stated.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

This type of information is not applicable to this study. The "ground truth" here is the biological variation of FT4 and TSH levels in healthy pediatric populations. Reference intervals are established statistically from a cohort of presumably healthy individuals, not through expert consensus on individual cases.

4. Adjudication Method for the Test Set

This is not applicable. There was no need for adjudication as the study focused on quantitative measurements and statistical analysis of those measurements to define reference ranges.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

This is not applicable. The device is an in-vitro diagnostic test for quantitative measurement, not an imaging device or a diagnostic aid requiring human interpretation that benefits from AI assistance. Therefore, an MRMC study and effect size for human improvement are not relevant.

6. Standalone Performance

Yes, this study inherently focuses on the standalone performance of the Dimension Vista® FT4 and TSH assays. The goal was to determine the measured analyte levels in a specific population (pediatrics) using the device alone, to establish reference intervals. There is no human-in-the-loop component for the determination of the numeric values.

7. Type of Ground Truth Used

The ground truth used to establish the reference intervals is biological reference data from a population of presumably healthy subjects within specified age groups. This involves statistical methods (non-parametric for children/adolescents, robust symmetric for infants, sometimes after log transformation for skewed data) to determine the central 95% of values.

8. Sample Size for the Training Set

This study is not framed in terms of a "training set" for an algorithm. The samples of 421 pediatric patients (82 infants, 191 children, 148 adolescents) directly served as the dataset from which the reference intervals were derived using statistical methods, which could be considered the "training" for the reference interval definition.

9. How the Ground Truth for the Training Set Was Established

The "ground truth" (i.e., the reference intervals) for the pediatric populations was established through statistical analysis of quantitative measurements of FT4 and TSH in serum/plasma samples from the 421 pediatric subjects.

• CLSI C28-A3c Guideline: The study followed the "Defining, Establishing and Verifying Reference Intervals in the Clinical Laboratory; Approved Guideline -- Third Edition (CLSI C28-A3c)" standard.
• Non-parametric approach: Used for children and adolescents, where the 2.5th and 97.5th percentiles of the distribution of values were calculated. This technique is typically used when the sample size is sufficient (usually >120).
• Robust symmetrical method (Horn and Pesce, CLSI EP28-A3c): Used for infants due to a sample size of less than 120 (82 patients).
• Log transformation: For the TSH assay in infants, the data was highly positively skewed, so a natural log transformation was applied to achieve a normal distribution before applying the robust symmetric method.
• Confirmation with Simulated Reference Intervals: Both methods were confirmed with two simulated reference intervals (non-parametric and robust symmetric) estimated as the means of the 95% upper and lower bounds from 1000 data sets.

Ask a specific question about this device

The Elecsys FT4 II Assay is for the in vitro quantitative determination of free Thyroxine in human serum and plasma. Measurements obtained by this device are used in the diagnosis and treatment of thyroid diseases.

The electrochemiluminescence immunoassay "ECLIA" is intended for use on Elecsys and cobas e immunoassay analyzers.

Elecsys FT4 II CalSet is used for calibrating the quantitative Elecsys FT4 II assay on the Elecsys and cobas e immunoassay analyzers.

1. The Elecsys F4 II Assay is a quantitative test for determination of free thyroxine in human serum and plasma. The total duration of the assay is 18 minutes. Elecsys FT4 II is a two-step competitive immunoassay with streptavidin microparticles, T4-specific polyclonal anti-T4-antibody (sheep) labeled with a sulfonyl-ruthenium complex, and electrochemiluminescence detection. Results are determined via a calibration curve which is instrument-specifically generated by a two-point calibration and a master curve provided via the reagent barcode.

(2) The Elecsys FT4 II CalSet is a ready-for-use buffer/protein (bovine serum albumin) matrix with added L-Thyroxine in two concentration ranges. FT4 II Cal1: 2 bottles, each containing 1.0 mL of calibrator 1 FT4 II Cal2: 2 bottles, each containing 1.0 mL of calibrator 2 Note: The reagent and calibrator are packaged separately.

Here's a breakdown of the acceptance criteria and the study details for the Elecsys FT4 II Assay, based on the provided 510(k) summary:

Acceptance Criteria and Device Performance for Elecsys FT4 II Assay

1. Table of Acceptance Criteria and Reported Device Performance

• Concentrations ≤ 0.4 ng/dL: SD ≤ 0.02 ng/dL
• Concentrations > 0.4-7.77 ng/dL: CV ≤ 5%
  Intermediate (Within-laboratory):
• Concentrations ≤ 0.4 ng/dL: SD ≤ 0.03 ng/dL
• Concentrations > 0.4-7.77 ng/dL: CV ≤ 8% | Repeatability:
• HS 1 (0.138 ng/dL): 4.0% CV (SD 0.006 ng/dL)
• HS 2 (1.03 ng/dL): 1.3% CV (SD 0.013 ng/dL)
• HS 3 (1.90 ng/dL): 1.3% CV (SD 0.024 ng/dL)
• HS 4 (4.93 ng/dL): 1.7% CV (SD 0.082 ng/dL)
• HS 5 (7.09 ng/dL): 1.8% CV (SD 0.127 ng/dL)
  Intermediate:
• HS 1 (0.138 ng/dL): 7.6% CV (SD 0.011 ng/dL)
• HS 2 (1.03 ng/dL): 2.3% CV (SD 0.023 ng/dL)
• HS 3 (1.90 ng/dL): 2.1% CV (SD 0.040 ng/dL)
• HS 4 (4.93 ng/dL): 3.3% CV (SD 0.163 ng/dL)
• HS 5 (7.09 ng/dL): 4.5% CV (SD 0.319 ng/dL)
  (All results meet specifications) |
  | Limit of Blank (LoB) | Claim will be set to ≤ 0.03 ng/dL. | LoB claim will be set to ≤ 0.03 ng/dL. (Study determined this value) |
  | Limit of Detection (LoD)| Claim will be set to ≤ 0.05 ng/dL. | LoD claim will be set to ≤ 0.05 ng/dL. (Study determined this value) |
  | Limit of Quantitation (LoQ) / Functional Sensitivity | Interassay coefficient of variation ≤ 20%. | Lot MP02: 0.059 ng/dL
  Lot P2: 0.069 ng/dL
  Lot P3: 0.080 ng/dL.
  Claim will be set to 0.1 ng/dL. (All lots show functional sensitivity below the claimed LoQ) |
  | Linearity | Significance level for deviation to higher order polynomial: 5%.
  Limits for deviation of higher order polynomial regression:
• 0.1 - 0.388 ng/dL: ±0.078 ng/dL

• 0.388 - 7.77 ng/dL: ± 10%
  Repeatability for linearity:

• 0.1 - 0.388 ng/dL: ±0.039 ng/dL

• 0.388 - 7.77 ng/dL: ± 5% | Linearity was confirmed in the range from 0.096 to 8.20 ng/dL. (Meets acceptance criteria) |
  | Exogenous Interferences - Drugs | Recovery of 100 ± 10% of the reference value (unspiked sample). | All compounds tested, except Levothyroxine and Furosemide, were found to be non-interfering. Levothyroxine and Furosemide will be listed as interfering substances. (Meets criteria for most, with noted exceptions) |
  | Exogenous Interferences - Anticoagulants | For regression analysis:

• Slope: 0.9 - 1.1
• Intercept: 0.3 – 7.77 ng/dL: 100 ± 10 % | No direct results given, but the study description implies successful demonstration up to 28 days for using Day 0 calibration. (Implies criteria were met) |
  | Reagent On-Board Stability | Recovery of samples compared to Day 1:
• LoD to 0.3 ng/dL: ± 0.05 ng/dL

• 0.3 to 7.77 ng/dL: 100±10% | Reagent kits can be stored on board for up to 28 days, and for 56 days with alternative storage (max 120 hours onboard). A new calibration is recommended every 7 days. (Implies criteria were met within these guidelines) |
  | Reagent Accelerated Stability | Recovery of samples compared to Day 0:

• LoD to 0.3 ng/dL: ± 0.05 ng/dL

• 0.3 to 7.77 ng/dL: 100 ± 10% | No direct results given, but used to support a 12-month shelf life claim. (Implies criteria were met for 3 weeks at 35°C, extrapolating to 12 months shelf life) |
  | Reagent Real-Time Stability | Recovery of 90-110% of the reference value. | In ongoing study, data for 0, 7, 10, 13, 16, 19, and 25 months will be available. Supports a 12-month shelf life claim based on both accelerated and initial real-time data. Study will continue for 24 months. (On-going, current data supports 12 months) |
  | Reagent Stability after first opening (2-8°C) | Recovery of samples compared to Day 0:

• ≤ 0.3 ng/dL: ± 0.05 ng/dL

• 0.3 - 7.77 ng/dL: 100 ± 10 % | Supports 84 days (12 weeks) when stored at 2-8ºC. (Implies criteria were met up to 85 days in the study) |
  | Reference Range Validation | No more than 6 (10%) of the 60 tested subjects should fall outside of the established reference range of 0.93-1.7 ng/dL. | 2 of 60 subjects fell outside the established reference range (0.93-1.7 ng/dL). Acceptance criteria met. The reference range can be transferred. (Meets acceptance criteria) |

2. Sample Sizes Used for the Test Set and Data Provenance

The document describes several performance studies, each with its own sample size:

• Precision (CLSI EP5-A2):
  • Sample Size: 84 determinations for each of 5 human serum samples (HS 1-5) and 2 PreciControl Universal (PCU 1-2). Tested in 2 replicates per sample/control per run, 2 runs per day for 21 days.
  • Data Provenance: Human sera (HS) and PreciControl Universal (PCU). No explicit country of origin is stated, but standard CLSI guidelines imply well-characterized, often commercially sourced, samples. The study is prospective in nature for assessing device performance.
• Limit of Blank (LoB) (CLSI EP17-A2):
  • Sample Size: 60 determinations of one analyte-free human serum sample (5-fold determination in each run, 1-2 runs/day for 4 days).
  • Data Provenance: Analyte-free human serum (T4 depleted).
• Limit of Detection (LoD) (CLSI EP17-A2):
  • Sample Size: Single measurement per run for four days on five low-level human serum samples. Total determinations not explicitly summed, but would be 5 samples * 1 measurement * (1-2 runs/day) * 4 days.
  • Data Provenance: Low-level human serum samples.
• Limit of Quantitation (LoQ) / Functional Sensitivity (CLSI EP17-A2):
  • Sample Size: 8 low-level human serum samples tested in single replicates for four days, 1-2 runs per day. Total determinations not explicitly summed.
  • Data Provenance: Native human serum samples diluted with analyte-free human serum matrix.
• Linearity (CLSI EP6-A):
  • Sample Size: 3 high analyte serum sample pools. For each pool, 13 concentrations (11 dilutions) across the measuring range were prepared and assayed in 3-fold determination within a single run.
  • Data Provenance: Human serum samples (spiked with L-Thyroxine) and fT4 depleted human serum.
• Specificity/Cross Reactivity:
  • Sample Size: Not explicitly stated, but compounds were tested in duplicate.
  • Data Provenance: Native human serum samples or human serum samples spiked with L-Thyroxine (single donors) spiked with potential cross-reactant compounds.
• Exogenous Interferences - Drugs:
  • Sample Size: 2 human serum samples approximately 1.1 ng/dL and 2.6 ng/dL of fT4. Each was spiked with 17 pharmaceutical compounds and 12 thyroid drugs/Furosemide. Spiked aliquots tested in triplicate, reference aliquot in 6-fold determination.
  • Data Provenance: Human serum samples (pools spiked with L-Thyroxine).
• Exogenous Interferences - Anticoagulants:
  • Sample Size: Between 53 and 63 serum/plasma pairs per sample material (presumably per type of anticoagulant). Tested in single determination.
  • Data Provenance: Native samples or samples spiked with L-Thyroxine (single donors) drawn into various primary tubes (serum, Li-Heparin, K2-EDTA, K3-EDTA-plasma, Li-Heparin Plasma Separation Tubes).
• Endogenous Interferences:
  • Sample Size: 3 serum samples (low, mid, high fT4 concentrations) for each interfering substance.
  • Data Provenance: Pooled human serum samples spiked with L-Thyroxine.
• Method Comparison (vs. predicate):
  • Sample Size: 170 human serum samples. Out of these, 11 were spiked with analyte, and 4 diluted with FT4 free serum.
  • Data Provenance: Human serum obtained from commercial vendors or remnant clinical samples.
• Calibration Stability:
  • Sample Size: 5 human serum samples (spiked with L-Thyroxine) and 2 controls. Each tested in duplicate.
  • Data Provenance: Human serum samples and controls.
• Reagent On-Board Stability:
  • Sample Size: Not explicitly stated, but refers to testing "samples" and controls at various time points.
  • Data Provenance: Human serum samples and controls.
• Reagent Accelerated Stability:
  • Sample Size: 5 human serum samples (spiked with L-Thyroxine) and 2 controls. Each tested in two-fold determination.
  • Data Provenance: Human serum samples and controls.
• Reagent Real-Time Stability:
  • Sample Size: PreciControl Universal 1 and PreciControl Universal 2 tested in duplicate at various time points (0, 7, 10, 13, 16, 19, 25 months).
  • Data Provenance: PreciControl Universal 1 and 2 (controls).
• Reagent Stability after first opening (2-8°C):
  • Sample Size: 5 human serum samples (spiked with L-Thyroxine) and 2 controls. Each tested in two-fold determination.
  • Data Provenance: Human serum samples and controls.
• Reference Range Validation Study (CLSI C28-A3c):
  • Sample Size: 60 subjects (30 males and 30 females).
  • Data Provenance: Subjects with normal TSH values (presumed healthy individuals). The study is prospective for reference range validation.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

This type of immunoassay (Elecsys FT4 II Assay) is a quantitative chemical measurement device. The "ground truth" for its performance is established through comparison to established analytical methods and reference standards, rather than expert interpretation of images or clinical findings.

• For studies like Linearity, Specificity, Interferences, and Analytical Sensitivity, ground truth samples are typically prepared with known concentrations of analytes or interferents, or are compared against validated reference methods.
• For Method Comparison, the ground truth is established by comparing the new device's results against a legally marketed predicate device (Elecsys FT4 Assay, K961489), which itself would have been validated against reference methods or clinical outcomes.
• For the Reference Range Validation Study, the "ground truth" for subject selection was having "normal TSH values" as measured by the Elecsys TSH assay, identifying a healthy population.

Therefore, the concept of "experts establishing ground truth" in the traditional sense (e.g., radiologists reviewing images) is not directly applicable here. The ground truth is inherent in the design of the analytical studies and the use of calibrated standards and reference methods.

4. Adjudication Method for the Test Set

Not applicable. As described above, this is a quantitative chemical measurement device; results are numerical and not subject to human interpretation or adjudication in the way a diagnostic imaging study would be. Deviations from expected values or comparisons to reference methods would be evaluated statistically.

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

Not applicable. This is not a device involving human readers interpreting cases or AI assistance.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was Done

Yes, the device (Elecsys FT4 II Assay) functions as a standalone, automated immunoassay without human intervention in the measurement process itself. All performance studies described (Precision, LoB, LoD, LoQ, Linearity, Specificity, Interferences, Method Comparison, Stability studies) evaluate the algorithm's performance and the device's analytical characteristics directly.

7. The Type of Ground Truth Used (Expert Consensus, Pathology, Outcomes Data, etc.)

The ground truth used across the various studies for this in vitro diagnostic device includes:

• Known concentrations: For studies like LoB, LoD, LoQ, Linearity, Specificity, and Interferences, samples are often prepared with precisely known concentrations of the analyte or interfering substances.
• Predicate device comparison: For method comparison, the results are compared to the legally marketed Elecsys FT4 Assay (K961489), which serves as the established analytical "truth" for substantial equivalence.
• Biological/clinical normality: For the Reference Range Validation Study, the ground truth for inclusion criteria was "normal TSH values" in healthy individuals.
• Calibration standards: The assays rely on accurate calibration using the Elecsys FT4 II CalSet, which is itself traceable to established standards (Enzymun-Test, standardized using equilibrium dialysis for FT4).

8. The Sample Size for the Training Set

Not applicable in the typical sense of machine learning. This is an immunoassay, not an AI or machine learning algorithm that requires a "training set" to learn patterns. The "training" of such a system would involve optimizing assay reagents and parameters during development, based on extensive R&D, but not a distinct "training set" of patient data as understood in AI/ML contexts.

9. How the Ground Truth for the Training Set Was Established

Not applicable, as there isn't a "training set" in the context of an immunoassay. The development and optimization of the assay would rely on:

• Analytical chemistry principles: Designing the competitive immunoassay with specific antibodies and detection systems.
• Biochemical characterization: Ensuring the specificity of T4-specific antibodies.
• Manufacturing controls: Producing reagents with consistent quality and concentration.
• Calibration curve development: Establishing the relationship between signal (electrochemiluminescence) and analyte concentration using a predefined set of calibrators with known concentrations that are traceable to reference methods (e.g., equilibrium dialysis).

Ask a specific question about this device