FDA 510(k) Search - By Innolitics (SaMD and AI Experts)

Immunoassay for the in vitro quantitative determination of cardiac troponin T (cTnT) in lithium heparin plasma. The immunoassay is intended to aid in the diagnosis of myocardial infarction.

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

Device Description

The Elecsys Troponin T Gen 5 STAT Immunoassay is a one-step sandwich immunoassay on the cobas e 601 and cobas e 801 analyzer. The assay uses streptavidin-coated microparticles, a biotinylated monoclonal anti-cardiac Troponin T-specific antibody, a monoclonal anti-cardiac Troponin T-specific antibody labeled with a ruthenium complex and electrochemiluminescence detection. Results are determined using a calibration curve that is generated specifically on each instrument by a 2-point calibration and a master curve (6-point calibration) provided with the reagent bar code.

AI/ML Overview

Here's an analysis of the acceptance criteria and supporting study for the Elecsys Troponin T Gen 5 device, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The document primarily focuses on demonstrating substantial equivalence to a predicate device rather than explicitly stating individual acceptance criteria for each analytical performance metric. However, for some metrics, "predetermined acceptance criterion" is mentioned. I will extract those with stated criteria and their outcomes.

Metric	Acceptance Criteria (where stated)	Reported Device Performance
Precision (Repeatability)	Not explicitly stated.	Site 1: CVs for samples 1-5 range from 0.9% to 3.3%.
Site 2: CVs for samples 1-5 range from 2.0% to 3.3%.
Site 3: CVs for samples 1-5 range from 1.1% to 3.0%.
Combined 21-day study (Table 1.1): CVs for samples 1-6 range from 1.4% to 2.3%.
Precision (Intermediate)	Not explicitly stated.	Site 1: CVs for samples 1-5 range from 1.3% to 3.3%.
Site 2: CVs for samples 1-5 range from 2.0% to 4.1%.
Site 3: CVs for samples 1-5 range from 3.3% to 4.8%.
Combined 21-day study (Table 1.1): CVs for samples 1-6 range from 2.9% to 4.2%.
Precision (Reproducibility)	Not explicitly stated.	Combined Sites (Table 1.3): CVs for samples 1-5 range from 3.0% to 8.8%.
Limit of Blank (LoB)	≤ 2.5 ng/L	All lots met the criterion. The LoB claim in labeling is 2.5 ng/L.
Limit of Detection (LoD)	≤ 3 ng/L	All lots met the criterion. The LoD claim in labeling is 3 ng/L.
Limit of Quantitation (LoQ)	20% CV (intermediate precision) goal, acceptable at 6 ng/L	All lots met the predetermined acceptance criterion of 6 ng/L. The LoQ claim in labeling is 6 ng/L.
Linearity	Deviation from linearity within specifications.	The linear range is reported as 6 to 13766 ng/L. The deviation from linearity was within specifications.
High Dose Hook Effect	Not explicitly stated (implied: no hook effect within expected range)	No High Dose Hook effect was seen up to 111326 ng/L.
HAMA Interference	Maximum of 10% deviation from expected concentration.	Specifications were met for two native Li-Heparin plasma samples. A maximum of 10% deviation was seen for a HAMA concentration of 644 ug/L (80% of dose spiked). Labeling reports testing of 322 ug/L.
Endogenous Interference (Biotin)	Biotin interference within specifications (implied by "No interference was seen up to 1200 ng/mL").	Biotin concentrations up to 1200 ng/mL showed "No interference seen". At higher concentrations (e.g., 2500 ng/mL, 3600 ng/mL), significant negative bias was observed, which suggests concentrations above 1200 ng/mL are problematic.
Other Endogenous Interferents	No interference seen.	Intralipid (2000 mg/dL), Bilirubin (66.0 mg/dL), Hemoglobin (200 mg/dL), Rheumatic Factor (1200 IU/mL), Human Serum Albumin (7.00 g/dL), Cholesterol (310 mg/dL) showed "No interference seen".
Cross-Reactivity	Specifications met.	Skeletal muscle TnT (30 ng/mL), Skeletal muscle TnI (100 ng/mL), Cardiac TnI (12.5 ng/mL), Human TnC (100 ng/mL) showed "No interference seen up to" the listed concentration, and "Specifications were met".
Exogenous Interference (Drugs)	Specifications met.	Seventeen common pharmaceutical compounds and 22 cardiac specific drugs showed "specifications were met".

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

Analytical Performance Studies (Precision, LoB, LoD, LoQ, Linearity, Interference, Cross-Reactivity):
- Sample Types: Human Li-Heparin plasma samples (native, pooled, or spiked with recombinant Troponin T).
- Provenance: Not explicitly stated for specific samples, but implied to be from internal or external sites where the tests were conducted (e.g., "one internal site," "three different sites (one internal and two external sites)"). These are typically retrospective samples used for method validation.
- Sample Sizes (examples):
  - Precision (Repeatability/Intermediate): Six human samples (sample pools) for 21 days with two replicates/day (4 replicates total per day for 21 days). Total 6 samples * 4 replicates/day * 21 days = 504 measurements per sample material.
  - Precision (Reproducibility): Five human Li-Heparin plasma samples and two controls, measured five-fold for 5 days on three analyzers at three sites. Total 7 materials * 5 replicates * 5 days * 3 sites = 525 measurements across all sites.
  - LoB: One analyte-free Li-Heparin plasma sample, measured in ten-fold determinations in each of six runs over at least three days. Total 60 measurements.
  - LoD: Five low-analyte Li-Heparin plasma samples, measured in two-fold determination, six runs over at least three days. Total 60 determinations per reagent lot (so 180 for 3 lots).
  - LoQ: Ten native Li-Heparin plasma pools and two controls, collected over 21 days with two runs per day. Total 12 materials * 2 runs/day * 21 days = 504 runs, with variance components calculated from these.
  - Linearity: Three high analyte Li-Heparin plasma samples diluted with three low analyte samples to generate at least sixteen concentrations (15 dilutions). Assayed in 3-fold determinations within a single run.
  - HAMA/Endogenous Interference: Two samples with low cTnT spiked with HAMA (in duplicate); Three native Li-Heparin plasma samples (low, medium, high cTnT) for endogenous interferents (3-fold determination).
  - Cross-Reactivity: Three concentrations of troponin samples spiked with four different cross-reactants, tested in three-fold determination.
  - Exogenous Interference (Drugs): Native Li-Heparin plasma samples spiked with 17 common pharmaceutical compounds and 22 cardiac drugs (3-fold determination).
Clinical Performance Studies (APACE Trial and second multicenter study):
- APACE Trial:
  - N: 1074 subjects enrolled, 3023 available samples. 188 adjudicated diagnoses of MI.
  - Provenance: International, multicenter prospective trial.
- Second Multicenter Study:
  - N: 1679 subjects presenting emergently with chest pain were enrolled. 173 adjudicated MIs. 1675 subjects evaluated on the cobas e 601 analyzer.
  - Provenance: Multicenter study (locations not specified, but typically clinical sites).

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

APACE Trial: "Independent adjudication committee which included cardiologists."
Second Multicenter Study: "Independent adjudication committee which included cardiologists and emergency medicine physicians."
Qualifications: "Cardiologists" and "emergency medicine physicians" are the stated qualifications. Specific years of experience are not provided.

4. Adjudication Method for the Test Set

APACE Trial: "In the case of a disagreement, a third independent cardiologist was used as the tie breaker." This describes a 2+1 or 3-way consensus adjudication method.
Second Multicenter Study: "Final diagnoses were determined by an independent adjudication committee... using the universal guidelines." The specific tie-breaking method is not explicitly stated, but implies a consensus process.

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

No, an MRMC comparative effectiveness study was not done. The document describes a method comparison study (comparing the candidate device to a predicate device) and clinical performance studies (diagnostic sensitivity/specificity of the device against a clinical ground truth).
There is no mention of human readers' performance with and without AI assistance, nor is there an effect size for human improvement with AI. This is an in vitro diagnostic device, not a diagnostic imaging AI algorithm that would typically involve human reader studies.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance)

Yes, this is a standalone performance study. The Elecsys Troponin T Gen 5 immunoassay is an in vitro diagnostic device that provides quantitative measurements of cardiac troponin T (cTnT). Its performance (analytical and clinical) is evaluated against established methods and clinical endpoints, independent of real-time human interpretation or intervention in the measurement process itself. The "human-in-the-loop" would be the clinician interpreting the result, but the device's measurement itself is standalone.

7. Type of Ground Truth Used

Clinical Performance Studies (APACE Trial & second multicenter study):
- Adjudicated Clinical Diagnosis of Myocardial Infarction (MI). This ground truth was established by an independent adjudication committee (cardiologists, sometimes with emergency medicine physicians) using established diagnostic criteria (e.g., ACC/ESC/AHA guidelines, including ECG changes, symptoms, and elevation of cardiac troponin). For the APACE trial, 60-day follow-up information was also used.
Analytical Performance Studies:
- For most analytical studies like precision, LoB, LoD, LoQ, linearity, and interference, the ground truth is often defined by:
  - Reference materials/standards: For calibration and standardization.
  - Known concentrations: For spiked samples and dilution series.
  - Analyte-free samples: For LoB determination.
  - Predicate device results: For method comparison studies. This isn't strictly "ground truth" but a reference for comparative equivalence.

8. Sample Size for the Training Set

The document describes a measurement device (immunoassay), not an AI/ML algorithm that is "trained" in the conventional sense on a dataset. Therefore, there isn't a "training set" in the context of an AI algorithm learning from data.
The device's calibration and standardization, however, rely on specific procedures:
- Calibration Curve: Generated specifically on each instrument by a 2-point calibration and a master curve (6-point calibration) provided with the reagent bar code. This essentially "calibrates" the device to provide accurate quantitative results.
- Standardization: Against Elecsys Troponin T STAT assay (4th generation), which in turn was standardized against the Enzymun-Test Troponin T (CARDIAC T) method. This refers to the traceable method of establishing the quantitative scale.

9. How the Ground Truth for the Training Set Was Established
As noted in point 8, this device does not utilize a "training set" in the context of AI/ML. Its operational "ground truth" or reference for quantitative measurement is established through:

Known Calibrator Values: The master curve calibration relies on calibrators with accurately assigned values, which are traceable back to a primary reference method (Enzymun-Test Troponin T, and further through the 4th generation assay).
Analytical Validation Studies: The various analytical studies (precision, linearity, LoB, LoD, LoQ) are validation steps to ensure the device accurately measures cTnT across its intended range, rather than a "training" phase. These studies confirm the device's ability to consistently provide results that align with expected or known values.

Summary

Regulation Number and Section

§ 862.1215 Creatine phosphokinase/creatine kinase or isoenzymes test system.

(a)
Identification. A creatine phosphokinase/creatine kinase or isoenzymes test system is a device intended to measure the activity of the enzyme creatine phosphokinase or its isoenzymes (a group of enzymes with similar biological activity) in plasma and serum. Measurements of creatine phosphokinase and its isoenzymes are used in the diagnosis and treatment of myocardial infarction and muscle diseases such as progressive, Duchenne-type muscular dystrophy.(b)
Classification. Class II.