Search Results

Immunoassay for the in vitro quantitative determination of thyroglobulin in human serum and plasma. Determination of Tg is used as an aid in monitoring for the presence of persistent or recurrent/metastatic disease in patients who have differentiated thyroid cancer (DTC) and have had thyroid surgery (with or without ablative therapy). The electrochemiluminescence immunoassay "ECLIA" is intended for use on cobas e immunoassay analyzers.

Device Description

The Tg II immunoassay makes use of a two-step, double antigen sandwich principle using a biotinylated monoclonal Tg-specific antibody and monoclonal Tg-specific antibodies labeled with a ruthenium complex. The Tg II immunoassay is intended for the in vitro quantitative determination of thyroglobulin in human serum and plasma. Determination of Tg is used to aid in monitoring for the presence of persistent or recurrent/metastatic disease in patients who have differentiated thyroid cancer (DTC) and have had thyroid surgery (with or without ablative therapy). It is intended for use on the cobas e immunoassay analyzers. Results are determined via a calibration curve which is instrument-specifically generated by 2-point calibration and a master curve provided via the reagent barcode or e-barcode.

AI/ML Overview

The Elecsys Tg II device is an immunoassay intended for the in vitro quantitative determination of thyroglobulin in human serum and plasma, used as an aid in monitoring for the presence of persistent or recurrent/metastatic disease in patients who have differentiated thyroid cancer (DTC) and have had thyroid surgery.

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

1. Table of Acceptance Criteria and Reported Device Performance

The provided document doesn't explicitly list "acceptance criteria" for all performance measures in a comparative table against the reported performance. However, based on the studies conducted and their results, one can infer the implicit acceptance criteria by observing the measured performance and statements like "All deviations from linearity met the specification" or "Non-significant interferences were defined as %interferences within ± 10 %".

Below is a table summarizing the reported device performance, with inferred acceptance criteria where direct ones are not explicitly stated, but are implied by the reported "met specifications" or similar statements.

Performance Characteristic	Inferred Acceptance Criteria	Reported Device Performance
Clinical Performance
Sensitivity	High sensitivity required for monitoring recurrent/metastatic disease (e.g., >90%).	98.91% (91/92) with 95% CI: (94.10%; 99.81%)
Specificity	Acceptable specificity for the intended use (the exact value is not explicitly stated as an initial acceptance criterion, but the reported value is presented as performance).	53.42% (234/438) with 95% CI: (48.74%; 58.05%)
Negative Predictive Value (NPV)	High NPV desired for ruling out disease (-ve result, truly -ve).	99.89% with 95% CI: (99.42%; 99.98%) (calculated at 4.99% prevalence)
Positive Predictive Value (PPV)	Acceptable PPV for the intended use.	10.03% with 95% CI: (9.16%; 11.03%) (calculated at 4.99% prevalence)
Analytical Performance
Limit of Blank (LoB)	Must be very low (e.g., in the picogram/mL range) to detect low levels of Tg. No explicit criterion given, but the reported value is the outcome of the study designed to determine it.	0.02 ng/mL (Determined according to CLSI EP17-A2)
Limit of Detection (LoD)	Must be very low, enabling early detection of disease recurrence. No explicit criterion given, but the reported value is the outcome of the study designed to determine it.	0.04 ng/mL (Determined according to CLSI EP17-A2)
Limit of Quantitation (LoQ)	%CV of within-laboratory precision ≤ 20% and %bias within ±15%.	0.1 ng/mL (%CV for samples around this level ranged from 7.56% to 4.00% for repeatability, and within-laboratory CVs were 9.34%, 8.75%, 5.67% for HS1, HS2, HS3 respectively, all below 20%. Bias not explicitly shown in summary table but met criteria.)
Linearity	Deviations from linearity ≤ ±10% for values ≥0.3 ng/mL and within ±0.03 ng/mL for values <0.3 ng/mL.	All deviations from linearity met the specification.
Measuring Range	To cover a broad range of clinically relevant Tg concentrations.	0.1 ng/mL – 500 ng/mL. Extended measuring interval is 500 – 5,000 ng/mL for 1:10 diluted samples.
High-Dose Hook Effect	No hook effect within clinically relevant high concentrations (e.g., up to 120,000 ng/mL or higher, implying reliable readings even at very high concentrations).	No hook effect observed up to ≥ 120,000 ng/mL Tg.
Matrix Comparison	Performance with Li-heparin, K2-EDTA, and K3-EDTA plasma should be similar to serum.	Matrix comparison studies showed that performance of Elecsys Tg II assay with these matrices are similar.
Biotin Interference	Non-significant interference, generally defined as %interference within ± 10%.	Biotin interference claim is set to 1200 ng/mL in labeling (meaning up to this concentration, interference is acceptably low, i.e., within ±10%).
Endogenous Substance and Pharmaceutical Interference	Non-significant interferences, defined as %interferences within ± 10 % at tested concentrations.	Non-significant interferences observed for Bilirubin, Hemoglobin, Intralipid, Biotin, IgG, Albumin, and 17 commonly used pharmaceuticals, and special drugs at tested concentrations (e.g., Iodide, Amiodarone, L-T4, etc.).
HAMA Interference	No significant HAMA interference (e.g., recovery within acceptable range, like ±10-20%).	No significant HAMA interference at 805 µg/L HAMA.
Cross-Reactivity	Cross-reactivity should be minimal (e.g., within a small percentage).	TSH: Within ±0.36% at 1000 mIU/L. TBG: Within ±0.0001% at 200000 ng/mL.
Stability
Shelf-life Stability	Reagent stable for the stated shelf-life (e.g., 15 months).	Data support a shelf-life of up to 15 months at 2-8°C.
Reagent After Opening	Reagent stable for a specified duration after first opening (e.g., 84 days) when stored at 2-8°C.	Up to 84 days (12 weeks) after opening, stored at 2-8°C.
On-board Reagent Stability	Reagent stable for a specified duration on the analyzer (e.g., 28 days).	Up to 28 days (4 weeks).
Sample Stability	Samples stable for specified durations under different storage conditions (e.g., 14 days at 2-8°C, 24 months at -20°C).	14 days at 2-8°C, 14 days at 15-25°C, and 24 months at -20 (± 5)°C for serum, Li-Heparin plasma, K2-EDTA and K3-EDTA plasma.

Study Information:

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

Clinical Performance Study (Combined Cohorts):
- Sample Size: 530 samples available for analysis.
- Data Provenance:
  - Origin: Collected from 9 sites across the U.S.
  - Nature:
    - Longitudinal Cohort: Serum samples collected from subjects 4-12 weeks post-thyroidectomy (with/without radioiodine ablation) and at 4 additional time points (6, 12, 18, 24 months post-surgery/ablation).
    - Cross-sectional Cohort: Samples collected from subjects with structural disease at a single visit. This cohort was used to increase the number of observations from patients with structural disease.
Reference Range Study:
- Sample Size: 244 healthy males and 219 healthy females (total 463).
- Data Provenance: Not explicitly stated (e.g., age range is given, but not country of origin, retrospective/prospective).
Expected Values in DTC post-thyroidectomy:
- Sample Size: 127 subjects with differentiated thyroid cancer (100 female; 27 male) with no evidence of disease for 4 or more years post-total/near total thyroidectomy.
- Data Provenance: Not explicitly stated.
Precision (Within-laboratory): 84 runs for each sample.
Reproducibility (Site-to-site): 146-149 runs for each sample.
LoB, LoD, LoQ, Linearity, Hook Effect, Interference, Cross-reactivity, Stability Studies: Varying number of samples/replicates as described in the respective sections (e.g., LoB/LoD: 5 native samples, 2 replicates/run, 6 runs over 4 days; LoQ: 7 low-level human serum samples, 5 replicates/run, 1 run/day over 5 days).

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

The document does not explicitly state the number of experts or their specific qualifications for establishing the ground truth of "structural disease" in the clinical study. However, the definition of structural disease is provided:

Ground Truth Definition: "Structural disease was defined as evidence of disease on ultrasound, cross sectional or functional imaging, or biopsy proven disease as determined by the investigator."
Implied Experts: This definition suggests that the "investigator" (likely a medical professional such as an endocrinologist, radiologist, or pathologist, depending on the type of evidence) determined the presence of structural disease. The number of such investigators involved in determining ground truth per patient or across the study is not specified.

4. Adjudication Method for the Test Set

The document does not explicitly describe an adjudication method (e.g., 2+1, 3+1) for establishing the ground truth of structural disease. It states that structural disease was "determined by the investigator." This implies a single determination rather than a consensus or adjudicated process among multiple experts for each case.

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

No, an MRMC comparative effectiveness study was not explicitly done. The study focuses on the standalone performance of the Elecsys Tg II assay rather than comparing human reader performance with and without AI assistance. The device itself is an immunoassay, not an AI imaging interpretation tool that would typically involve human readers.

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

Yes, the studies presented are primarily standalone (algorithm only) performance evaluations. The Elecsys Tg II is an automated immunoassay system, and its performance (sensitivity, specificity, precision, accuracy, linearity, stability, etc.) is measured directly on biological samples without active human interpretation assistance in the analytical or clinical performance reported here. The "results are determined via a calibration curve which is instrument-specifically generated."

7. The Type of Ground Truth Used

Clinical Performance Study: The ground truth for the clinical performance study (sensitivity, specificity, NPV, PPV) was based on structural disease (SD+ or SD-). This was determined by:
- Evidence of disease on ultrasound.
- Cross-sectional or functional imaging.
- Biopsy-proven disease.
- These were "determined by the investigator."
Analytical Studies: For analytical performance (LoB, LoD, LoQ, linearity, interference, stability), the ground truth generally involved:
- Reference standards/materials: e.g., CRM 457.
- Known analyte concentrations: Obtained by spiking or using characterized serum pools.
- Clinical laboratory standards: (e.g., CLSI guidelines).

8. The Sample Size for the Training Set

The document does not mention a separate "training set" for the Elecsys Tg II device in the context of machine learning or AI algorithms. As an immunoassay, the "training" analogous to an AI system would be the development and optimization of the assay reagents, protocols, and calibration curves. The provided information relates to the validation of the final product. The "calibration curve" is established for the assay, which is a form of internal "training" for the instrument's interpretation of signal to concentration.

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

As noted above, for an immunoassay, there isn't a "training set" in the typical AI sense. The development of the assay relies on establishing accurate measurements against recognized reference standards and materials (e.g., CRM 457, characterized serum pools with known thyroglobulin concentrations). These standards serve as the "ground truth" to ensure the assay accurately quantifies thyroglobulin. The assay is then calibrated using these standards to accurately relate the measured signal (chemiluminescent emission) to the thyroglobulin concentration in patient samples.

Ask a Question

Ask a specific question about this device

Page 1 of 1