(21 days)
Immunoassay for the in vitro quantitative determination of thyrotropin in human serum and plasma. Measurements of TSH are used in the diagnosis of thyroid and pituitary disorders.
The electrochemiluminescence immunoassay "ECLIA" is intended for use on cobas e immunoassay analyzers.
The Elecsys TSH immunoassay makes use of a sandwich test principle using monoclonal antibodies specifically directed against human TSH. The antibodies labeled with ruthenium complex) consist of a chimeric construct from human and mouse specific components. The Elecsys TSH immunoassay is used for the in vitro quantitative determination of thyroid stimulating hormone in human serum and plasma. It is intended for use on the cobas e immunoassay analyzers.
The Elecsys TSH device is an immunoassay for the in vitro quantitative determination of thyrotropin in human serum and plasma, used in the diagnosis of thyroid and pituitary disorders. It is an electrochemiluminescence immunoassay (ECLIA) intended for use on cobas e immunoassay analyzers. The key change in the updated device is a two-step approach to block biotin interference by adding an antibody to bind free biotin in the sample and changing the linker on the biotinylated capture antibody.
Here's an analysis of the acceptance criteria and the study that proves the device meets them:
1. A table of acceptance criteria and the reported device performance
The document provides performance data across various non-clinical studies. The acceptance criteria are generally implied by "All samples met the predetermined acceptance criterion" or "All lots met the predetermined acceptance criterion" for studies like precision, LoB, LoD, LoQ, and linearity. For interference studies, the "No interference seen up to" values represent the performance vs. a defined limit. For lot-to-lot reproducibility, the comparability of SDs and CVs implicitly confirms acceptance. For method comparison, the statistical results (slope, intercept, correlation coefficient, bias) are compared against internal acceptance criteria.
| Clinical / Technical Feature | Acceptance Criteria (Explicit or Implied) | Reported Device Performance |
|---|---|---|
| Repeatability & Intermediate Precision | All samples met the predetermined acceptance criterion. | CVs for repeatability ranged from 0.7% to 3.4%. CVs for intermediate precision ranged from 1.5% to 11.2%. |
| Lot-to-Lot Reproducibility | Calculated SDs and CVs for multiple lots comparable to single lot precision study. | Calculated SDs and CVs for multiple lots were comparable. |
| Limit of Blank (LoB) | All lots met the predetermined acceptance criterion. | 0.0025 µIU/mL |
| Limit of Detection (LoD) | All lots met the predetermined acceptance criterion. | 0.005 µIU/mL |
| Limit of Quantitation (LoQ) | All lots met the predetermined acceptance criterion. | 0.005 µIU/mL |
| Linearity/Assay Reportable Range | All deviations within predetermined acceptance criteria. | Linear in the range from 0.004 - 102 µIU/mL. |
| High Dose Hook Effect | No hook effect observed up to a specified concentration. | No hook effect up to 1000 µIU/mL TSH. |
| Biotin Interference (Endogenous) | Biotin interference not exceeding a specified threshold. | No biotin interference in serum concentrations up to 1200 ng/mL. (Previous limitation was ≤ 102 nmol/L or ≤ 25 ng/mL). |
| Lipemia (Intralipid) Interference | No interference seen up to 1500 mg/dL. | No interference seen up to 2000 mg/dL. |
| Hemoglobin Interference | No interference seen up to 1000 mg/dL. | No interference seen up to 1000 mg/dL. |
| Bilirubin Interference | No interference seen up to 41 mg/dL. | No interference seen up to 66 mg/dL. |
| Rheumatoid Factor (RF) Interference | No interference seen up to 1500 IU/mL. | No interference seen up to 1500 IU/mL. |
| Immunoglobulin (IgG) Interference | No interference seen up to 2 g/dL. | No interference seen up to 3.98 g/dL. |
| Immunoglobulin (IgM) Interference | No interference seen up to 0.5 g/dL. | No interference seen up to 0.72 g/dL. |
| Analytical Specificity (Cross-Reactivity) | All cross-reactivities met the predefined acceptance criterion at the specified concentration. | LH, FSH, hCG showed 0.000% cross-reactivity at high concentrations; hGH not detectable. |
| Exogenous Interferences (Drugs) | Each compound found to be non-interfering at the drug concentration. | All 30 tested drugs (commonly and specially used) showed no significant interference at concentrations at least 3x maximum daily doses (or 1x for some). |
| Sample Matrix Comparison | Regression analysis (Passing/Bablok) data consistent with acceptance criteria for various plasma types and different separating gels. | Slope (0.976 - 0.983), Intercept (-0.0006 - -0.021), Correlation (0.999 - 1.00) for serum vs. plasma. Recovery acceptable for PST/SST. |
| Method Comparison to Predicate | All data met predefined acceptance criteria for agreement between candidate (updated assay) and predicate (current assay). | Passing-Bablok: Slope 0.974, Intercept -0.0002, Correlation 0.999. Bias at 0.27 µIU/mL: -2.7%, Bias at 4.2 µIU/mL: -2.6%. |
| Stability | Pre-specified acceptance criteria were met. | Stability data supports Roche Diagnostic's claims as reported in package inserts. |
2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
- Precision (Repeatability & Intermediate Precision): The test set involved 2 replicates per run for 21 days, across 2 runs per day, for PreciControl Universal, PC Thyro Sensitive, and 5 human serum samples. This is a prospective study design for precision. Sample types were native, single human donors as well as pools.
- Lot-to-Lot Reproducibility: 2 replicates of each of 5 human serum samples per run, 2 runs per day, for 21 days (7 days per lot, n=28 determinations per lot). Prospective design. Sample types were native, single human donors as well as pools.
- Limit of Blank (LoB): Five blank samples with two replicates each per run, for 6 runs on ≥ 3 days (total 60 determinations for analyte free samples). Prospective design.
- Limit of Detection (LoD): Five low analyte samples with two replicates each per run, for 6 runs on ≥ 3 days (total 60 replicates per sample per reagent lot). Prospective design.
- Limit of Quantitation (LoQ): 25 replicates per sample per reagent lot, over 5 days (1 run per day). Prospective design.
- Linearity/Assay Reportable Range: Three high analyte human serum samples were diluted and measured in 3-fold determination within a single run. Prospective design.
- High Dose Hook Effect: Three human serum samples spiked with analyte, dilution series performed, measured in one-fold determination. Prospective design.
- Endogenous Interference: Varied by interferent. For Biotin, Lipemia, Hemoglobin, Bilirubin, RF, IgG, IgM, samples were spiked with interfering substances and diluted into a dilution pool in 10% increments. The number of individual samples/pools is not explicitly stated but implies multiple. Prospective design.
- Analytical Specificity/Cross-Reactivity: A native human serum sample pool was used for each potential cross-reacting compound. Prospective design.
- Exogenous Interferences (Drugs): Two human serum samples (native serum pools) were used. Prospective design.
- Sample Matrix Comparison: A minimum of 56 serum/plasma pairs per sample material (Li-heparin, K2-EDTA, K3-EDTA plasma) were tested in singleton. For PST/SST, blood from five donors was used, measured in duplicate. Prospective design.
- Method Comparison to Predicate: 138 samples (129 native human serum samples and 9 diluted human serum samples, single donors as well as pools diluted) were measured in singleton. Prospective design.
Data Provenance: The document does not explicitly state the country of origin for the human serum and plasma samples. However, the manufacturer, Roche Diagnostics, operates globally with establishments in Mannheim and Penzberg, Germany, and Indianapolis, USA. The studies typically indicate the use of "human serum" or "human serum samples" without further geographic specification. All described studies appear to be prospective experimental designs conducted in a laboratory setting for device validation.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)
This device is an in vitro diagnostic immunoassay testing for a quantifiable biomarker (TSH), not an imaging device or a device requiring expert interpretation of complex clinical data to establish ground truth for its performance characteristics. The ground truth for such assays is established through analytical methods and reference standards (e.g., spectrophotometry for linearity, spiked samples for interference, reference materials for precision, comparison to a predicate device). No human experts are used to "establish primary ground truth" in the sense of clinical diagnosis for these analytical performance studies. The ground truth is the actual concentration of the analyte, or the known characteristics of the samples (e.g., spiked amount of interferent).
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
Not applicable. As described above, this is an in vitro diagnostic analytical performance study, not a clinical study requiring adjudication of diagnoses or interpretations by experts.
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 device is an automated immunoassay, an in vitro diagnostic (IVD) test, not an imaging device or AI-driven diagnostic tool that would involve human "readers" or AI assistance.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Yes, the studies described are all standalone performance studies of the Elecsys TSH immunoassay system. It's an automated device (cobas e immunoassay analyzer) that provides quantitative results without human intervention in the measurement process itself, beyond sample loading and general operation/maintenance.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)
The ground truth used for these analytical studies includes:
- Known concentrations: For LoB, LoD, LoQ, and linearity, the samples are either analyte-free, at known low concentrations, or dilutions from known high concentrations.
- Spiked samples: For interference studies (endogenous and exogenous), known amounts of interfering substances are added to samples.
- Reference materials/standards: For precision, controls with defined concentrations are used. Traceability is to the 2nd IRP WHO Reference Standard 80/558.
- Comparison to a legally marketed predicate device: For method comparison, the results of the new device are compared quantitatively to those of the predicate device.
- Clinical samples (native human serum/plasma): These are used to assess the device's performance across a range of physiological concentrations and in real-world matrices for studies like precision, linearity, and matrix comparison.
8. The sample size for the training set
This document describes a 510(k) submission for a revised immunoassay, not a machine learning or AI-based device. Therefore, there is no "training set" in the context of algorithm development. The development of the assay itself would have involved extensive R&D and analytical testing to optimize reagents and protocols, but this is distinct from training an AI model on a dataset.
9. How the ground truth for the training set was established
Not applicable, as there is no "training set" in the AI/ML context for this device. The assay's performance characteristics are established through the non-clinical studies detailed in the summary.
§ 862.1690 Thyroid stimulating hormone test system.
(a)
Identification. A thyroid stimulating hormone test system is a device intended to measure thyroid stimulating hormone, also known as thyrotrophin and thyrotrophic hormone, in serum and plasma. Measurements of thyroid stimulating hormone produced by the anterior pituitary are used in the diagnosis of thyroid or pituitary disorders.(b)
Classification. Class II.