Search Results

The ARCHITECT 2nd Generation Testosterone assay is a chemiluminescent microparticle immunoassay (CMIA) for the quantitative determination of testosterone in human serum and plasma. Measurements of testosterone are used in the diagnosis and treatment of disorders involving the male sex hormones (androgens), including primary and secondary hypogonadism, delayed or precocious puberty, impotence in males and, in females, hirsutism (excessive hair) and virilization (masculinization) due to tumors, polycystic ovaries, and adrenogenital syndromes.

The ARCHITECT 2nd Generation Testosterone Calibrators are for the calibration of the ARCHITECT i System when used for the quantitative determination of testosterone in human serum and plasma.

The ARCHITECT 2nd Generation Testosterone Controls are for the verification of the accuracy and precision of the ARCHITECT i System when used for the quantitative determination of testosterone in human serum and plasma.

Device Description

The ARCHITECT 2nd Generation Testosterone Reagent Kit consists of 100 (1 x 100) or 400 (4 x 100) tests. Each kit contains 1 or 4 bottle(s) each of Microparticles. Conjugate. Assay Specific Diluent, and Specimen Diluent.

Microparticles 1 or 4 Bottle(s) (6.6 mL) Anti-Testosterone (sheep, monoclonal) coated . microparticles in BIS Tris buffer with protein (bovine) stabilizer. Minimum concentration: 0.1% solids. Preservative: ProClin 300.
. Conjugate -1 or 4 Bottle(s) (6.9 mL) Testosterone acridinium-labeled conjugate in BIS Tris buffer with surfactant stabilizer. Minimum concentration: 6.5 nmol/L. Preservative: ProClin 300.
. Assay Specific Diluent - 1 or 4 Bottle(s) (25.0 mL) Testosterone Assay Diluent containing phosphate and glycine in citrate buffer. Preservative: ProClin 300.
. Specimen Diluent - 1 or 4 Bottle(s) (12.2 mL) Testosterone Specimen Diluent containing PBS buffer. Preservative: ProClin 300.

Each ARCHITECT 2nd Generation Testosterone Calibrator Kit contains 6 Bottles (4.0 mL each) of ARCHITECT 2nd Generation Testosterone Calibrators A-F. Calibrator A contains PBS buffer. Calibrators B through F contain testosterone in PBS buffer. All calibrators contain a protein (bovine) stabilizer. Preservative: ProClin 300.

Each ARCHITECT 2nd Generation Testosterone Control Kit contains 3 Bottles (8.0 mL each) of ARCHITECT 2nd Generation Testosterone Controls. The Low, Medium, and High Controls contain testosterone in PBS buffer with a protein (bovine) stabilizer. Preservative: ProClin 300.

AI/ML Overview

Here's an analysis of the acceptance criteria and the study proving the device meets them, based on the provided text:

Device: ARCHITECT 2nd Generation Testosterone assay, Calibrators, and Controls.

1. A table of acceptance criteria and the reported device performance

Performance Characteristic	Acceptance Criteria	Reported Device Performance (ARCHITECT 2nd Generation Testosterone)
Precision	Within-laboratory (total) imprecision of ≤ 10% CV for samples with testosterone concentrations ≥ 14.4 ng/dL to 1009 ng/dL.	All reported within-laboratory (total) %CV values were ≤ 5.2%, well within the ≤ 10% CV acceptance criterion for the control levels (9.02-229.95 ng/dL) and the human serum panel (60.72-62.35 ng/dL) tested. Specific values range from 2.6% to 5.2%.
Limit of Quantitation (LoQ)	LoQ of ≤ 4.33 ng/dL with an inter-assay imprecision of < 20%.	Observed LoQ: 4.33 ng/dL with an inter-assay CV of 15.6%. This meets the acceptance criterion.
Limit of Blank (LoB)	Not explicitly stated as an acceptance criterion in this section, but a value is presented.	Observed LoB: 1.73 ng/dL.
Limit of Detection (LoD)	Not explicitly stated as an acceptance criterion in this section, but a value is presented.	Observed LoD: 2.67 ng/dL.
Measuring Interval (Reportable Range)	Measuring interval from 4.33 ng/dL (LoQ) to 1500.00 ng/dL.	Demonstrated range: 4.33 ng/dL (LoQ) to 1500.00 ng/dL. This matches the acceptance criterion.
Linearity	Linear range supporting the Measuring Interval (4.33 ng/dL to 1500 ng/dL), with an absolute deviation from linearity of 3.605 ng/dL for ≤ 14.4 ng/dL and 10% for > 14.4 ng/dL.	Demonstrated linear range: 3.82 - 1862.27 ng/dL. This supports the claimed measuring interval.
Interference (Endogenous Substances)	Potential interference from bilirubin, hemoglobin, total protein, triglycerides, biotin, and SHBG to be < 10%.	All tested endogenous substances showed % Interference < 10%. Specific values ranged from -7.0% to 4.4%.
Method Comparison (vs. LCMS)	Slope of 1.0 ± 0.2 and a correlation coefficient (r) of ≥ 0.95 for samples with testosterone concentrations ranging from 4.33 ng/dL (LoQ) to 1500.00 ng/dL.	Overall (n=138): Slope 1.00 (95% CI: 0.98, 1.03), r = 0.994. Meets criteria. Female Specimens (n=73): Slope 0.82 (95% CI: 0.77, 0.88), r = 0.985. Meets r criterion, but slope is at the lower end of the criterion. Male Specimens (n=65): Slope 1.10 (95% CI: 1.07, 1.13), r = 0.990. Meets criteria.

2. Sample sized used for the test set and the data provenance

Precision (Within-Laboratory): 80 replicates per control level/panel, instrument, and reagent lot combination. This means 4 (control/panel levels) * 2 (instruments) * 2 (reagent lots) = 16 combinations, so 16 * 80 = 1280 individual measurements for controls and 1 human serum panel.
- Provenance: "human serum panel" is mentioned, implying human origin. The study was conducted based on NCCLS document EP5-A2 guidance. Retrospective/prospective is not specified, but typically lab performance studies use carefully controlled, often prepared, samples.
Limit of Quantitation (LoQ), Limit of Blank (LoB), Limit of Detection (LoD): The sample size specific to the LoQ study design (NCCLS document EP17-A) is not explicitly stated, but the resulting LoQ was determined.
Linearity: The sample size for the linearity study (NCCLS document EP6-A) is not explicitly stated.
Interference (Endogenous Substances): Sample size for the interference study (CLSI protocol EP7-A2) is not explicitly stated for each substance, but evaluation was performed.
Interference (Potentially Interfering Drugs and Other Compounds): Sample size for this interference study (CLSI document EP7-A2) is not explicitly stated for each compound.
Tube Type Matrix Comparison: N = 54 samples for each evaluation tube type to compare against the control tube type.
Expected Values:
- Males (21-49 years): 129 samples
- Males (≥ 50 years): 71 samples
- Females (21-49 years): 129 samples
- Females (≥ 50 years): 52 samples
- Provenance: All from "apparently healthy individuals". Country of origin/retrospective/prospective is not specified.
Method Comparison (vs. LCMS): n = 138 total samples (73 female specimens, 65 male specimens).
- Provenance: Human samples are implied as "female specimens" and "male specimens" were analyzed. Retrospective/prospective is not specified.

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

This is an in vitro diagnostic (IVD) device for measuring a quantitative analyte (testosterone). The "ground truth" for such devices is typically established through reference methods or highly accurate analytical techniques, not through expert consensus in the way image-based diagnostics might.

For Method Comparison: Liquid Chromatography - Tandem Mass Spectrometry (LCMS) was used as the reference method. LCMS is a highly precise and accurate analytical technique, often considered a "gold standard" for quantitative biochemical measurements. The "experts" in this context would be the skilled laboratory personnel operating and validating the LCMS system. No specific number or qualifications of "experts" are provided in the document for the LCMS ground truth, as it's an analytical measurement rather than a subjective interpretation.

4. Adjudication method for the test set

Not applicable for this type of quantitative diagnostic device. Adjudication methods (like 2+1 or 3+1) are typically used in studies where human experts interpret results (e.g., radiology images) and a consensus is needed to establish ground truth. For an IVD, the ground truth is established by a reference analytical method (LCMS in this case).

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 an AI-assisted diagnostic device where human readers interact with AI. It is a standalone immunoassay measuring testosterone levels.

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

Yes, the studies presented demonstrate the standalone performance of the ARCHITECT 2nd Generation Testosterone assay. The device directly measures testosterone concentrations in samples, with performance characteristics (precision, sensitivity, linearity, interference, method comparison) evaluated solely based on the assay's output against established analytical standards and reference methods. There is no human interpretation or intervention in the generation of the primary testosterone measurement result.

7. The type of ground truth used

Analytical Standards/Known Concentrations: For precision, LoQ, LoB, LoD, linearity, and interference studies, the ground truth often comes from prepared samples with known, spiked concentrations of testosterone and/or interfering substances.
Reference Method: For the method comparison study, Liquid Chromatography - Tandem Mass Spectrometry (LCMS) was used as the ground truth.
Known Clinical Status: For "Expected Values" (reference ranges), the ground truth was based on samples from "apparently healthy individuals" categorized by age and sex. This is a clinical rather than an analytical ground truth.

8. The sample size for the training set

This document describes performance validation studies for an established immunoassay technology. Immunoassay development typically involves extensive optimization and characterization during the R&D phase, which would involve numerous samples. However, the document does not distinguish between a "training set" and a "test set" in the context of machine learning, as this is a chemical assay, not an AI model. The sample sizes listed in point #2 above pertain to the performance evaluation studies used to demonstrate the device meets its acceptance criteria. No separate "training set" for an algorithm is mentioned as there is no AI algorithm.

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

Not applicable as there is no training set mentioned in the context of an AI algorithm. For the development of the assay itself, the ground truth for optimizing antibodies and assay parameters would historically come from highly purified testosterone standards and well-characterized clinical samples, leveraging similar analytical reference methods (e.g., LCMS) to define true concentrations.

Ask a Question

Ask a specific question about this device

Page 1 of 1