Search Results
Found 2 results
510(k) Data Aggregation
(332 days)
The Axis-Shield HoloTC RIA is an in-vitro diagnostic assay for quantitative measurement of holotranscobalamin (vitamin B12 bound to transcobalamin) in human serum or plasma. Measurements obtained by this device are used in the diagnosis and treatment of vitamin B12 deficiency. HoloTC RIA is calibrated with HoloTC Calibrators. HoloTC controls are assayed for the verification of the accuracy and precision of the HoloTC RIA.
The Axis-Shield HoloTC RIA is a competitive binding immunoassay in which a specific monoclonal antibody is used to capture transcobalamin from the patient sample. Thereafter the procedure is as commonly used in vitamin B12 assays. The cobalamin (vitamin B12) is released from the transcobalamin using dithiothreitol and sodium hydroxide. The released cobalamin (vitamin B12) then competes for a limited amount of intrinsic factor with added 'Co labelled vitamin B12. The Axis-Shield HoloTC RIA differs from the predicate device in two main aspects:
-
- The use of a transcobalamin specific antibody, this allows the quantitation of only the cobalamin bound to the protein transcobalamin as opposed to measurement of cobalamin bound to all proteins in the predicate device and
-
- The detection signal is radioactivity (37Co) as opposed to chemiluminescence in the predicate device.
Here's an analysis of the provided text regarding the Axis-Shield HoloTC RIA device, addressing your requested information:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state pre-defined acceptance criteria for all performance metrics. However, where implicitly stated or inferable, they are included.
| Performance Characteristic | Acceptance Criteria (Implicit/Explicit) | Reported Device Performance |
|---|---|---|
| Method Comparison | Close linear correlation, ideally slope ≈ 1, intercept ≈ 0, high r² | HoloTC pmol/L = 0.55 * VB12-2 pmol/L; r² = 0.52 (compared to predicate Bayer Advia Centaur VB12 assay) |
| Reference Interval | n/a (established for diagnosis) | 37-171 pmol/L (95% central interval for Finnish population, n=303) |
| External Evaluation (CV) | <15% (for three levels of serum samples) | CVs for three serum samples were "within the given acceptance criteria, <15 %" |
| Precision (Within-run CV) | Not explicitly stated but implied to be acceptable | 11% (low), 5% (medium), 8% (high) |
| Precision (Total CV) | Not explicitly stated but implied to be acceptable | 11% (low), 6% (medium), 8% (high) |
| Linearity (Working Range) | Meet all criteria for linearity | Demonstrated within 10-160 pmol/L |
| Linearity (Recovery Test) | Within acceptance criteria | Demonstrated within acceptance criteria |
| Linearity (Dilution Correlation) | High r², slope ≈ 1, intercept ≈ 0 | r² = 0.998, slope = 1.01 ± 0.02, y-intercept = - 5 ± 2 pmol/L (for dilutions 13-182 pmol/L) |
| LOQ | Not explicitly stated/implied to be acceptable for clinical use | 8-176 pmol/L |
| LOD | Not explicitly stated/implied to be acceptable for clinical use | 6 pmol/L |
| Cross-reactivity (HC) | ±10% (for up to 70,000 pg/mL HC) | Well within acceptance criteria |
| Interference | <10% interference | <10% for bilirubin, hemoglobin, total protein, lipids |
| Serum/EDTA Plasma Comparison | No significant difference | No significant difference demonstrated |
| In-use Stability | Not explicitly stated/implied to be acceptable | Stability up to three months after opening |
| Concordance | Not explicitly stated/implied to be acceptable for diagnosis | 80% (weighted combined estimate) |
| Sensitivity | Not explicitly stated/implied to be acceptable for diagnosis | 99.5% (weighted combined estimate) |
| Specificity | Not explicitly stated/implied to be acceptable for diagnosis | 76.3% (weighted combined estimate) |
2. Sample Size Used for the Test Set and Data Provenance
- Method Comparison: 392 patient samples. Data provenance not explicitly stated (e.g., country of origin, retrospective/prospective).
- Reference Interval: 303 individuals from a "Finnish population of normal individuals." This suggests prospective collection for establishing normative data within that specific population.
- External Evaluation: Three different levels of native and pooled serum samples, in addition to kit controls Low and High. Number of samples not given, but tested at two external study sites (UK and Denmark).
- Precision Studies: Three levels of serum samples (low, medium, high). Number of samples for each level not explicitly stated, but duplicate measurements were performed.
- Linearity (Dilution): High HoloTC sample diluted into six dilutions. The original sample size is not stated.
- Serum/EDTA Plasma Comparison: 50 paired serum and EDTA plasma samples from the same donors.
- Sensitivity and Specificity: Samples from 3 study sites: 112 classified as "likely vitamin B12 deficient" and 313 as "not likely vitamin B12 deficient." Total of 425 samples. Data provenance (country of origin, retrospective/prospective) not explicitly stated.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
- The document does not mention the use of experts to establish ground truth for the test set.
- For sensitivity and specificity, samples were classified based on "cobalamin and MMA levels," indicating an objective biochemical measure rather than expert interpretation of a diagnostic image or clinical presentation.
4. Adjudication Method for the Test Set
- No adjudication method is mentioned, as there is no indication of expert interpretation requiring consensus. The ground truth for classification of B12 deficiency was based on biochemical markers (cobalamin and MMA levels).
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done
- No, an MRMC comparative effectiveness study was not done. This device is an in-vitro diagnostic assay which provides a quantitative measurement, not an imaging device requiring human interpretation, therefore, the concept of a multi-reader study or human readers improving with AI assistance is not applicable. The comparison was against a predicate assay (Bayer Advia Centaur VB12 assay).
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
- Yes, the entire study focuses on the standalone performance of the Axis-Shield HoloTC RIA assay (which can be considered the "algorithm only" in the context of an IVD) in measuring holotranscobalamin levels. As an in-vitro diagnostic quantitative assay, its performance is inherently standalone, providing a numerical result.
7. The type of ground truth used
- Method Comparison: The ground truth for comparison was the measurements obtained from the predicate device, the Bayer Advia Centaur VB12 assay.
- Reference Interval: The ground truth was based on the measured HoloTC levels in a population of "normal individuals."
- Sensitivity and Specificity: The ground truth for classifying samples as "likely vitamin B12 deficient" or "not likely vitamin B12 deficient" was based on cobalamin and MMA levels. This is a form of outcomes data or established biochemical markers for the condition.
8. The Sample Size for the Training Set
- The document does not explicitly mention a "training set" in the context of an AI/ML model development. For an immunoassay like the HoloTC RIA, the "training" involves the development and optimization of the assay reagents, protocols, and calibration curves. The provided numbers are for validation or evaluation sets.
9. How the Ground Truth for the Training Set Was Established
- As there is no explicit mention of a "training set" for an AI/ML model, this question is not directly applicable. For the development of the immunoassay itself, the "ground truth" during development would be established through a rigorous process of biochemical characterization, optimization using known concentrations of analytes, and comparisons to established reference methods or highly purified standards to ensure accurate and specific binding. This is an engineering/biochemical development process rather than establishing a "ground truth" for a dataset in the AI/ML sense.
Ask a specific question about this device
(176 days)
Ask a specific question about this device
Page 1 of 1