Search Results

The IMMULITE® 2000 TSI (thyroid-stimulating immunoglobulins) Assay is an in vitro diagnostic immunoassay for the semi-quantitative determination of thyroid stimulating autoantibodies specific to thyroid stimulating hormone receptors (TSHR) in human serum (including Serum Separator tubes) or plasma (K2-EDTA or lithium heparin). The IMMULITE® 2000 TSI Assay is for use on the IMMULITE® 2000 system. The measurement of thyroid stimulating autoantibodies, in conjunction with other clinical and laboratory findings, is used as an aid in the diagnosis of patients suspected of having Graves' disease.

The IMMULITE® TSI Calibration Verification Material (CVM) is for in vitro diagnostic use in the verification of calibration of the IMMULITE® TSI Assay on the IMMULITE® 2000 Systems.

The IMMULITE 2000 TSI assay kit consists of the following components:

• TSI bead pack coated with MAb (3D7) anti-TSHR anchor antibody and hTSHR Capture Chimera
• TSI reagent wedge containing hTSHR-Chimera alkaline phosphatase conjugate
• TSI adjustors: low and high, containing TSI negative heat-inactivated bovine serum and thyroid stimulating human MAb (M22)
• TSI controls: negative, low, and high, containing TSI negative human serum and thyroid stimulating MAb (M22)
• Multi-Diluent 2

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are generally implied from the "Drift Specifications" for stability and the statistical measures for precision and method comparison. The device performance is the "reported performance".

• Repeatability %CV: Ranged from 3.5% to 7.0%.
• Within Lab %CV: Ranged from 5.0% to 8.3%. |
  | Linearity/Reportable Range | Linearity data for % Difference to be ± 15% or 0.50 IU/L (whichever is greater). | Linearity data for % Difference was shown to be ± 15% or 0.50 IU/L (whichever is greater) for most samples; tested within range 0.50 - 40.0 IU/L. |
  | Assay/Component Stability Drift | Reagents and Beads:
• Control 1 at ≤ -60°C: ≤ 15% from Day 0 mean
• Control 2 at ≤ -60°C: ≤ 10% from Day 0 mean
• CVM 1 at ≤ -60°C: ≤ 0.15 IU/L from Day 0 mean
• MDP 1: ≤ 20% from Day 0 mean
• MDP 4: ≤ 10% from Day 0 mean
• Cal J: ≤ 10% from Day 0 mean
  (Also, all Control results must be within QC established range to validate the run). | Stability Claims Achieved:
• Kit, unopened: 12 Months (2-8 °C)
• Bead Pack, open: 90 Days (2-8 °C)
• Reagent wedge, open and on-board: 90 Days (2-8 °C)
• Sample diluent, open: 30 Days (2-8 °C)
• Sample diluent, open frozen aliquotted: 6 Months (-20 °C)
• Adjustors open: 90 Days (2-8 °C)
• Adjustors frozen aliquotted: 4 Months (-20 °C)
• Controls open: 90 Days (2-8 °C)
• Controls frozen aliquotted: 6 Months (-20 °C)
• CVM, unopened: 12 Months (2-8 °C)
• CVM, opened and reconstituted: 30 Days (2-8 °C)
  Sample Stability: 24 hours at 20-25°C, 7 days at 2-8°C, and 12 months at -20°C for serum and plasma. |
  | Limit of Blank (LoB) | LoB should be low enough for clinical utility. | Highest LoB by lot was determined to be 0.03 IU/L. |
  | Limit of Detection (LoD) | LoD should be low enough for clinical utility. | Highest LoD by lot was determined to be 0.06 IU/L. |
  | Analytical Specificity (Interference) | No interference (≤ 10% different than control sample). | Interferents (Intralipid, Hemoglobin, Bilirubin, K2-EDTA): No interference (≤ 10% difference) except for hemoglobin (≥ 200mg/dL potentially affects recovery) and short draw K2-EDTA (may result in under-recovery).
  HAMA: Individual sample bias of 90% for positive, negative, and overall). | Positive Agreement: 95.8% (95% CI: 93.0 – 97.7)
  Negative Agreement: 87.7% (95% CI: 84.5 - 90.5)
  Overall Agreement: 91.0% (95% CI: 88.8 - 92.9) |
  | Matrix Comparison (Regression Coefficient) | Strong correlation (e.g., R-value > 0.95 or 0.98). | SST Serum: Slope 1.01, Intercept 0.00, Correlation Coefficient 0.99
  K2-EDTA Plasma: Slope 1.03, Intercept -0.01, Correlation Coefficient 0.99
  Lithium Heparin Plasma: Slope 0.99, Intercept -0.01, Correlation Coefficient 0.99 |
  | Clinical Sensitivity and Specificity (at 0.55 IU/L cut-off) | High clinical sensitivity and specificity (generally >95%). | Clinical Sensitivity: 98.6% (95% CI: 96.8 – 99.5)
  Clinical Specificity: 98.5% (95% CI: 96.8 – 99.5) |
  | Reference Range (97.5th percentile) | Clinically appropriate range. | 180 measurements). Data provenance not specified.

• LoD: 5 TSI serum samples, tested in 4 replicates using 3 reagent kit lots for 3 days on 3 systems (total 36 observations per sample). Data provenance not specified.
• Analytical Specificity (Interference): Three serum pools (for endogenous interferents), HAMA positive serum, RF positive serum, and six potential biological interferent samples. Data provenance not specified.
• Method Comparison: 811 serum samples from patients with Graves' disease, other thyroid or autoantibody diseases. Specimens tested at two external sites and one internal site for the IMMULITE 2000 TSI assay, and at one external site for the Thyretain device. Data provenance not explicitly stated, but implies mixed sources (internal/external, likely retrospective patient samples).
• Matrix Comparison: Graves' Disease sets of matched serum and plasma samples (clot tube, lithium heparin, SST, K2-EDTA). Sample size not specified. Data provenance not specified.
• Clinical Sensitivity and Specificity: 361 treated and untreated hyperthyroid Graves' disease patients and 404 individuals with other thyroid or autoimmune diseases. Data provenance not specified (likely retrospective clinical samples).
• Reference Range: 842 serum samples from apparently healthy males (n=151), non-pregnant females (n=155), first trimester (n=169), second trimester (n=191), and third trimester (n=176) pregnant donors. Data provenance not specified.

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)

The document does not specify the number or qualifications of experts used to establish the ground truth for any of the test sets. For example, for clinical sensitivity and specificity, it refers to "Graves' Disease Diagnosis" and "other thyroid or autoimmune diseases" but does not detail how these diagnoses were definitively established or by whom. Similarly, for the method comparison, it refers to patient samples for Graves' disease or other thyroid/autoantibody diseases without detailing expert adjudication of these conditions.

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

The document does not describe any specific adjudication method for establishing ground truth diagnoses for the patient samples used in the studies. The diagnoses are simply stated as existing, implying they were pre-determined or established through typical clinical practice, rather than through a separate expert adjudication process for the study.

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

No, a multi-reader multi-case (MRMC) comparative effectiveness study was not done. This is an immunoassay device, not an imaging AI device that involves human readers interpreting cases. The comparison study was a "Method Comparison" between the new Immulite 2000 TSI assay and a predicate immunoassay device (Thyretain TSI Reporter BioAssay), analyzing patient samples.

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

Yes, the studies presented are all standalone validations of the Immulite 2000 TSI assay system. It is an automated immunoassay for in vitro diagnostic use, meaning it provides results directly from patient samples without human interpretation of images or other complex data requiring "human-in-the-loop" decision-making. The performance characteristics (precision, linearity, detection limits, clinical sensitivity/specificity, etc.) reflect the algorithm/device's performance alone.

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

The ground truth for the clinical studies (clinical sensitivity/specificity, method comparison) appears to be clinical diagnosis of Graves' disease or other thyroid/autoimmune diseases. For analytical performance studies (precision, linearity, LoB, LoD, interference), the ground truth is based on spiking known concentrations, using negative samples, or established reference standards like the NIBSC standard.

8. The sample size for the training set

The document does not explicitly mention a "training set" in the context of machine learning. This is a traditional immunoassay, not an AI device that typically involves distinct training and test sets in the same manner. The studies describe validation sets, not training sets for model development.

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

As there is no distinct "training set" in the context of an AI/machine learning model as understood in typical AI/ML submissions, this question is not directly applicable. For traditional assays, calibrators are used to establish the measurement and are themselves calibrated against reference materials (e.g., NIBSC standard 08/204 for the Immulite 2000 TSI assay).

Ask a specific question about this device