(360 days)
The Atellica® CI Analyzer is an automated, integrated system in vitro diagnostic tests on clinical specimens. The system is intended for the qualitative analysis of various body fluids, using photometry, turbidimetric, chemiluminescent, and integrated ionselective electrode technology for clinical use.
The Atellica® IM Thyroid Stimulating Hormone 3-Ultra (TSH3-UL) assay is for in vitro diagnostic use in the quantitative determination of thyroid-stimulating hormone (TSH, thyrotropin) in human serum and plasma (EDTA and lithium heparin) using the Atellica® CI Analyzer. Measurements of thyroid stimulating hormone produced by the anterior pituitary are used in the diagnosis of thyroid or pituitary disorders.
The Atellica® CH Albumin BCP (AlbP) assay is for in vitro diagnostic use in the quantitative measurement of albumin in human serum and plasma (lithium heparin, potassum EDTA) using the Atellica® CI Analyzer. Albumin measurements are used in the diagnosis and treatment of numerous diseases primarily involving the liver or kidneys.
The Atellica® CI Analyzer is an automated, integrated system designed to perform in vitro diagnostic tests on clinical specimens. The system is intended for the qualitative and quantitative analysis of various body fluids, using photometric, turbidimetric, chemiluminescent, and integrated ionselective electrode technology for clinical use.
The Atellica CI Analyzer with Atellica® Rack Handler supports both clinical chemistry (CH) and Immunoassay (IM) features and contains all the necessary hardware, electronics, and software to automatically process samples and generate results, including sample and reagent dispensing, mixing, and incubating.
The Atellica IM TSH3-UL assay is a third-generation assay that employs anti-FITC monoclonal antibody covalently bound to paramagnetic particles, an FITC-labeled anti-TSH capture mouse monoclonal antibody, and a tracer consisting of a proprietary acridinium ester and an anti-TSH mouse monoclonal antibody conjugated to bovine serum albumin (BSA) for chemiluminescent detection
The Atellica CH Albumin BCP (AlbP) assay is an adaptation of the bromocresol purple dy-e binding method reported by Carter and Louderback et al. In the Atellica CH AlbP assay, serum or plasma albumin quantitatively binds to BCP to form an albumin-BCP complex that is measured as an endpoint reaction at 596/694 nm coenzyme NAD+ functions only with the bacterial (Leuconostoc mesenteroides) enzyme employed in the assay.
The document provided is a 510(k) summary for in vitro diagnostic devices (IVDs), specifically the Atellica® CI Analyzer and its associated assays for Thyroid Stimulating Hormone (TSH3-UL) and Albumin (AlbP). IVDs, by their nature, measure specific analytes in biological samples and are evaluated against performance criteria such as precision, accuracy, linearity, and interference, rather than diagnostic accuracy metrics like sensitivity and specificity that would typically apply to AI/ML software. Therefore, many of the requested elements pertaining to AI/ML acceptance criteria and human-in-the-loop studies are not applicable to this type of device.
Here's a breakdown of the relevant information provided:
1. A table of acceptance criteria and the reported device performance:
The document describes the performance characteristics for the Atellica IM Thyroid Stimulating Hormone 3-Ultra (TSH3-UL) assay and the Atellica CH Albumin BCP (AlbP) assay. These are performance criteria, which serve as the acceptance criteria for the device's analytical performance.
Atellica IM Thyroid Stimulating Hormone 3-Ultra (TSH3-UL) Assay:
| Performance Characteristic | Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|---|
| Limit of Blank (LoB) | Must meet defined statistical criteria (CLSI EP17-A2.18) | 0.004 µIU/mL (mIU/L) |
| Limit of Detection (LoD) | Must meet defined statistical criteria (CLSI EP17-A2.18) | 0.008 µIU/mL (mIU/L) |
| Limit of Quantitation (LoQ) | Within-laboratory CV ≤ 20% | 0.008 µIU/mL (mIU/L) |
| Precision (Serum Samples) | Repeatability and Within-Laboratory CVs within acceptable ranges | Ranges from 1.1% to 1.5% for CV (Repeatability) and 1.9% to 3.3% for CV (Within-Laboratory) across various concentrations. |
| Assay Comparison (Serum) | Correlation coefficient (r) > 0.960 (per AlbP section, assumed similar for TSH3-UL) | r = 0.996 (compared to Atellica IM Analyzer) |
| Interfering Substances | Bias due to interfering substances ≤ 10% (for specific concentrations) | Hemoglobin, Bilirubin (conjugated/unconjugated), Lipemia (Intralipid®) show biases of -0.1% to -3%. |
| Other Substances | Bias due to these substances ≤ 10% (at specified TSH concentrations) | No interference (bias ≤ 10%) from listed substances (e.g., Biotin, Cholesterol, Acetaminophen, etc.) at tested concentrations. |
| Specimen Equivalency | Correlation coefficient (r) indicative of equivalence | Plasma (Lithium heparin) vs. Serum: r = 1.00; Plasma (EDTA) vs. Serum: r = 1.00 |
| High-Dose Hook Effect | Report > 150.000 µIU/mL (mIU/L) for high TSH concentrations | Samples with TSH concentrations as high as 3000 µIU/mL (mIU/L) will report > 150.000 µIU/mL (mIU/L). |
| Cross-Reactivity | Bias due to cross-reacting substances ≤ 5% | Human Chorionic Gonadotropin, Follicle Stimulating Hormone, Luteinizing Hormone show differences of -2.1% to 1.7%. |
| Onboard Dilution Recovery | Recovery within an acceptable range (e.g., 90-110%) | Mean recovery of 99.3% and 100.1% for serum, 100.5% and 99.3% for plasma across dilutions. |
| Linearity | Demonstrated linearity over the claimed measuring range (0.008-150.000 µIU/mL) | Y=0.9945*X-0.0011, demonstrating linearity. |
Atellica CH Albumin BCP (AlbP) Assay:
| Performance Characteristic | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Blank (LoB) | ≤ 0.1 g/dL (≤ 1 g/L) | 0.1 g/dL (1 g/L) |
| Limit of Detection (LoD) | ≤ 0.6 g/dL (≤ 6 g/L) | 0.5 g/dL (5 g/L) |
| Limit of Quantitation (LoQ) | Within-laboratory precision ≤ 10% | 0.5 g/dL (5 g/L) |
| Precision (Serum Samples) | Repeatability and Within-Laboratory CVs within acceptable ranges | Ranges from 0.6% to 1.3% for CV (Repeatability) and 1.7% to 2.6% for CV (Within-Laboratory) across various concentrations. |
| Reproducibility | Repeatability, Between-Day, Between-Instrument, Between-Lot, Total Reproducibility within acceptable ranges | Total Reproducibility CVs range from 1.4% to 1.9%. |
| Assay Comparison | Correlation coefficient (r) > 0.960 and slope 1.00 ± 0.10 | r = 0.999; y = 0.98x + 0.0 g/dL (compared to Atellica CH Analyzer) |
| Specimen Equivalency | Correlation coefficient (r) indicative of equivalence | Plasma (Lithium heparin) vs. Serum: r = 0.995; Plasma (Potassium EDTA) vs. Serum: r = 0.997 |
| Hemolysis, Icterus, Lipemia (HIL) | ≤ 10% interference from hemoglobin, bilirubin, and lipemia | Biases typically within 9% for tested concentrations. |
| Non-Interfering Substances | Bias due to these substances ≤ 10% | Biases typically within 10% for listed substances. |
| Linearity | Demonstrated linearity over the claimed measuring range (0.5-8.0 g/dL) | Y=0.9984*X+0.2891, demonstrating linearity. |
2. Sample sizes used for the test set and the data provenance:
- TSH3-UL Assay:
- Precision: 80 samples for each type (Serum A-F, EDTA Plasma A-C, Heparin Plasma A-C, Control 1-3).
- Assay Comparison (Serum): 112 samples.
- Interferences (Specific substances): Not explicitly stated how many samples per substance, but concentrations tested at two analyte levels.
- Specimen Equivalency: 64 samples for Plasma (Lithium heparin) and 64 for Plasma (EDTA).
- Onboard Dilution Recovery: 3 samples (Serum and Plasma) tested at two dilution levels.
- Linearity: Not explicitly stated, but "at least 14 levels created by mixing high and low serum samples" with N=5 replicates per level.
- AlbP Assay:
- LoD: 486 determinations (270 blank, 216 low level replicates).
- LoQ: n=5 replicates using 3 reagent lots over 5 days.
- Precision: N ≥ 80 for each sample (Serum 1-3, Serum QC 1).
- Reproducibility: 225 samples for each serum level (assayed n=5 in 1 run for 5 days using 3 instruments and 3 reagent lots).
- Assay Comparison (Serum): 106 samples.
- Specimen Equivalency: 76 samples for Plasma (Lithium heparin) and 55 for Plasma (Potassium EDTA).
- HIL: Not explicitly stated how many samples per interferent, but concentrations tested at two analyte levels.
- Non-Interfering Substances: Not explicitly stated how many samples per substance, but tested at two analyte concentrations.
- Linearity: "at least nine levels created by mixing the high and low pools of serum" with N=5 replicates per level.
Data Provenance: The document does not explicitly state the country of origin or whether the data was retrospective or prospective. Given it's a 510(k) submission for a medical device intended for broad use, it's highly likely the studies were prospective analytical validation studies conducted under controlled laboratory conditions, typically in multiple sites to ensure robustness.
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 information is not applicable to this type of device. The "ground truth" for clinical laboratory assays like TSH and Albumin comes from established analytical methods, reference materials, and accepted scientific principles of chemistry and immunology. It's about measuring the concentration of an analyte, not interpreting an image or diagnosing a condition based on expert consensus. The "experts" involved would be clinical chemists, laboratory scientists, and engineers responsible for assay development and validation, following established guidelines like those from CLSI (Clinical and Laboratory Standards Institute).
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set:
This is not applicable for this type of device. Adjudication methods are used in studies involving subjective interpretations (e.g., image reading) where multiple readers provide opinions that need to be reconciled to establish ground truth. For quantitative chemical assays, the "truth" is determined by reference methods and the intrinsic properties of the analyte, not by human consensus or adjudication.
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:
This is not applicable. An MRMC study is designed for evaluating the impact of a system on human readers' diagnostic performance, typically in the context of imaging. This document describes an automated in vitro diagnostic analyzer and its assays, which do not involve human "readers" in the sense of interpreting outputs like medical images.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
The performance characteristics presented (precision, linearity, assay comparison, interference, etc.) represent the standalone performance of the device and its assays. The Atellica® CI Analyzer and its assays are automated systems designed to perform measurements without human interpretative input beyond setting up the instrument and following standard laboratory procedures for running samples and quality control.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc):
The ground truth for these quantitative assays is established through:
- Reference Methods / Comparability: The performance is evaluated by comparing the new device's results to a legally marketed predicate device (Siemens Trinidad systems) which serve as the reference. This establishes the equivalence of the new device to already accepted technology.
- Traceability to International Standards: For TSH3-UL, traceability is to the World Health Organization (WHO) 3rd International Standard for human TSH (IRP 81/565). For AlbP, traceability is to ERM-DA470k Reference Material. These international standards or reference materials provide the "true" or accepted values against which the device's measurements are calibrated and verified.
- Analytical Procedures: The "ground truth" for characteristics like limit of detection, precision, and linearity are determined by rigorous statistical methods and established protocols (e.g., CLSI guidelines EP05-A3, EP07-ed3, EP09c-ed3, EP17-A2, EP06-ED2) during analytical validation.
8. The sample size for the training set:
This information is not applicable in the context of an IVD where "training set" implies machine learning or AI model development. For an IVD, there is a development and validation process. The number of samples for analytical validation studies (which is what is presented) is given under point 2.
9. How the ground truth for the training set was established:
As this is not an AI/ML device, the concept of a "training set" for an algorithm and its associated ground truth establishment methods (e.g., expert annotations) are not applicable. The "ground truth" or reference for the development and validation of these IVD assays is based on established laboratory practices, chemical principles, certified reference materials, and comparison to predicate devices, as described in point 7.
§ 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.