(52 days)
The Albumin BCP assay is an in vitro diagnostic test used for the determination of albumin in human serum or plasma. Albumin measurements are used in the diagnosis and treatment of numerous diseases primarily involving the liver or kidneys.
The assay is intended for professional use only.
For In Vitro Diagnostic use only.
Albumin BCP reagent is ready to use liquid reagent that is supplied in two configurations: fill volume 20 mL in a 20 mL wedge or 50 mL in a 50 mL wedge, 6 wedges/kit.
Here's the breakdown of the acceptance criteria and study information for the Albumin BCP device, based on the provided text:
Acceptance Criteria and Device Performance
| Study | Acceptance Criteria (Required Performance) | Reported Device Performance (Achieved) |
|---|---|---|
| Limit of Blank (LoB) | ≤ 1 g/L | 0.3 g/L (highest observed) |
| Limit of Detection (LoD) | ≤ 3 g/L | 0.8 g/L (highest observed) |
| Limit of Quantitation (LoQ) | ≤ 5 g/L | 1.5 g/L (claimed) |
| Precision | ≤ 2.5% CV (across all tested concentrations) | Highest %CV: 2.2% |
| Intra Assay Precision | ≤ 1.5% CV (across all tested concentrations) | All samples gave %CV lower than 1.5% (e.g., 0.40% to 0.95%) |
| Linearity (Measuring Range) | Absolute bias: - 2 g/L to + 2 g/L OR Relative bias: -6% to + 6% | Linear up to 70 g/L (e.g., y = 0.00 + 1.000x, r = 0.999) |
| Endogenous Interferences | % bias: ±10% for Hemoglobin (2000 mg/dL), Unconjugated bilirubin (66 mg/dL), Conjugated bilirubin (66 mg/dL), Lipids (as Triglycerides) (2000 mg/dL) | Met acceptance criteria for all tested substances at specified concentrations (Lipids up to 1200 mg/dL) |
| Reagent Stability | % Bias: within ± 10% vs initial measurement | Min %bias: -1.3%, Max %bias: 10.0% |
| Method Comparison | Regression slope of 1.00 (± 0.10) and a correlation coefficient (r) of ≥ 0.975 | Passing & Bablok: y = 0.94x + 1.01, r = 0.992; Linear fit: y = 0.95x + 0.78, r = 0.992 |
| Matrix Comparison (Serum vs. Lithium-Heparin plasma) | Regression slope of 1.00 (± 0.10) and a correlation coefficient (r) of ≥ 0.975 | Passing & Bablok: y = 1.01x - 0.34, r = 0.995; Linear fit: y = 1.00x + 0.01, r = 0.995 |
| Matrix Comparison (Serum vs. Potassium EDTA plasma) | Regression slope of 1.00 (± 0.10) and a correlation coefficient (r) of ≥ 0.975 | Passing & Bablok: y = 1.00x - 0.20, r = 0.996; Linear fit: y = 0.995x - 0.05, r = 0.996 |
Note regarding "Test Set" and "Training Set" terminology: For in vitro diagnostic assays measuring specific analytes, the concepts of "test set" and "training set" (as typically used in machine learning or image analysis) are not directly applicable in the same way. Instead, performance studies use different sample types (e.g., control materials, patient samples, spiked samples) to validate the analytical performance characteristics. The following answers reflect this distinction.
Study Details:
-
Sample sizes used for the test set and the data provenance:
- Limit of Blank (LoB): Not explicitly stated, but performed with three different reagent lots (F0390, F0391, F0480) and one calibrator lot (E0179). LoB is typically determined using replicate measurements of blank samples.
- Limit of Detection (LoD): Not explicitly stated, but inferred to be similar to LoB determination as it also uses reagent and calibrator lots.
- Limit of Quantitation (LoQ): Not explicitly stated, inferred to be similar to LoB/LoD determination.
- Precision Study: 80 replicates per level for three different reagent lots (F0390, F0391, F0480) across three levels (26.26-26.62 g/L, 40.53-40.67 g/L, 49.96-50.47 g/L) using human serum. An additional lot (90228) used 88 replicates per level (19.10 g/L, 40.18 g/L, 51.33 g/L).
- Intra Assay Precision Study: 20 replicates per level for three different reagent lots (F0390, F0391, F0480) across three levels (21.4-21.5 g/L, 35.6-35.8 g/L, 50.2-50.3 g/L) using human serum.
- Linearity (Measuring Range): Three different reagent lots (F0390, F0391, F0480) were tested across specified ranges (e.g., 4.17 to 78.30 g/L). Number of distinct samples within these ranges not explicitly stated.
- Endogenous Interferences Study: "2 aliquots of serum pool were prepared (Base and Test pool)" for two albumin concentrations (~35 g/L and ~50 g/L), with the test pool divided into 4 sub-aliquots and diluted. Specific number of interference samples not stated, but covered a range of dilution levels (100% down to 0%).
- Reagent Stability: Four different lots (F0390, F0391, F0480, 90228) were evaluated across three different concentration levels.
- Method Comparison: 128 serum samples, including 8 altered samples, covering the measuring interval 6.0 - 70 g/L.
- Matrix Comparison: 77 paired plasma/serum samples, including 7 altered samples, covering the assay's range, for both Lithium-Heparin plasma and Potassium EDTA plasma.
Data Provenance: The studies used human serum and plasma samples. The document does not explicitly state the country of origin of the data or whether the samples were collected retrospectively or prospectively. Given the context of a medical device submission, these would typically be from clinical laboratory settings.
-
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 is an in vitro diagnostic (IVD) assay for measuring a biochemical analyte (albumin). The "ground truth" for such assays is established by the reference methods or highly characterized materials used to calibrate and validate the assay. It does not involve human experts interpreting images or diagnosing conditions, but rather relies on established analytical standards and predicate devices.
-
Adjudication method (e.g. 2+1, 3+1, none) for the test set:
Not applicable. Adjudication methods like 2+1 or 3+1 are used for subjective interpretations (e.g., image reading) where disagreement among experts might arise. For quantitative IVD assays, performance is assessed against defined analytical criteria and reference values.
-
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 an in vitro diagnostic device, not an AI-assisted diagnostic tool that would involve human readers.
-
If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
Yes, the performance studies described are for the standalone functioning of the Albumin BCP assay on the AU680 Automatic Analyzer. This is inherent to the nature of an in vitro diagnostic test, where the device performs the measurement independently.
-
The type of ground truth used (expert concensus, pathology, outcomes data, etc):
The "ground truth" for this IVD device is established through:
- Reference materials and calibrators: Used to ensure accuracy and traceability of measurements (e.g., ERM-DA 470k/IFCC for standardization).
- Predicate device measurements: The method comparison study used a legally marketed predicate device (Siemens ADVIA 2400, ADVIA® Chemistry Albumin BCP assay) as a comparative standard.
- Analytical standards: Performance is measured against accepted analytical performance guidelines (e.g., CLSI documents EP17-A2, EP05-A3, EP15-A3, EP6-A, EP07-A2, EP09-A3) which define acceptable limits for various performance characteristics.
- Known concentrations: For studies like LoB, LoD, LoQ, Precision, and Intra-Assay Precision, samples with known or characterized concentrations (e.g., control materials, spiked samples, serum pools) are used to assess the device's accuracy and reproducibility.
-
The sample size for the training set:
Not applicable in the machine learning sense. The device is a chemical assay, not an algorithm trained on a dataset. Its analytical characteristics are inherently designed and validated through laboratory studies.
-
How the ground truth for the training set was established:
Not applicable. As explained above, this device does not utilize a "training set" in the context of machine learning. The analytical methods and performance targets are established through scientific principles of chemistry and validated using established laboratory practices and reference standards.
§ 862.1035 Albumin test system.
(a)
Identification. An albumin test system is a device intended to measure the albumin concentration in serum and plasma. Albumin measurements are used in the diagnosis and treatment of numerous diseases involving primarily the liver or kidneys.(b)
Classification. Class II.