(106 days)
In vitro test for the quantitative immunological determination of C-reactive protein in human serum and plasma on COBAS INTEGRA systems.
In vitro test for the quantitative determination of C-reactive protein in human serum and plasma on Roche/Hitachi cobas c systems.
Measurements of C-reactive protein aids in evaluation of the amount of injury to body tissues.
The C-Reactive Protein (Latex) assay is a particle enhanced turbidimetric assay. Human CRP agglutinates with latex particle coated with monoclonal anti-CRP antibodies. The precipitate is determined turbidimetrically at 552 nm (546 nm on cobas c 501 and c 311 analyzers).
Here's an analysis of the provided text regarding the acceptance criteria and the study that proves the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
Based on the provided text, the "acceptance criteria" are implied by the performance characteristics presented for the modified device, and in comparison to the predicate device. The submission focuses on demonstrating that the modified device's performance is either "Same" as the predicate or improved/updated without compromising safety or effectiveness.
| Feature | Predicate Device (K981897) Performance | Modified Device (C-Reactive Protein (Latex)) Performance | Comparison / Acceptance Implication |
|---|---|---|---|
| Precision | Level 1: CV within-run 1.8%, CV Total 2.9% | ||
| Level 2: CV within-run 1.5%, CV Total 2.7% | Same as Predicate | The modified device maintains the same precision as the predicate, indicating it continues to meet the established performance standard for precision. | |
| Lower Detection Limit | 0.25 mg/L | 1.00 mg/L* (*Performance has not changed. The specification is now stated as the claim) | This is a clarification/re-statement of the specification, not a change in performance. The acceptance criterion is effectively "performance has not changed from the predicate's 0.25 mg/L, but it is now claimed as 1.00 mg/L". This implies the device still meets the previous performance level even with a different stated claim. |
| Method Comparison (K981897 vs. K951595) | n=244 | ||
| r=0.993 | |||
| Lin. Regression: y=1.07x - 6.2 mg/L | |||
| P/B Regression: y=1.00x - 2.7 mg/L | |||
| Values: 0.62 to 421 mg/L | - | This section describes the predicate's original method comparison. The current submission's comparison (see next row) would be the new acceptance criteria for method comparison for the modified device. | |
| Method Comparison (Modified Device vs. Predicate) | - | n=150 | |
| r=0.999 | |||
| Lin. Regression: y=0.996x - 0.60 mg/L | |||
| P/B Regression: y=0.992x - 0.16 mg/L | |||
| Values: 0.62 to 362 mg/L | The acceptance criteria here would be indicated by the strong correlation coefficient (r=0.999) and regression equations close to y=x, demonstrating substantial equivalence to the predicate device. This shows the modified device provides comparable results to the predicate across the tested range. | ||
| Lipemia (L-index) Interference | Turbid samples exceeding 0.1 Absorbance recognized by "High Activity" check. Effective results after post-dilution. Triglyceride levels above 7.5 g/L decrease apparent CRP value significantly. | COBAS INTEGRA 400/400 plus: No significant interference up to L index of 1500 (lower concentration) and 623 (higher concentration). | |
| COBAS INTEGRA 700/800: No significant interference up to L index of 1094 (lower concentration) and 797 (higher concentration). | |||
| Poor correlation between L index and triglycerides. Turbid samples exceeding 0.1 Absorbance recognized by "High Activity" check; effective results after post-dilution. | The modified device provides more specific and quantitative data on Lipemia interference using Intralipid. The acceptance criteria are that, up to specified L-index values at different concentration ranges on different instruments, there is "No significant interference." This demonstrates improved characterization (and thus understanding of limitations) for interference. The overall performance (turbid sample recognition and post-dilution) is "Same." | ||
| Measuring Range | 0-160 mg/L (0-1600 mg/L with postdilution, factor 10) | Integra 700, 800, 400, 400 plus, cobas c 111: 1-200 mg/L (1:10 dilution for higher) | |
| cobas c 501 / c 311: 1-250 mg/L (1:3 dilution for higher) | The acceptance criteria are the expanded measuring ranges (up to 200 mg/L or 250 mg/L depending on the instrument) and the demonstrated efficacy of the specified post-dilution protocols (1:10 or 1:3). This shows an improvement in the device's capabilities without compromising accuracy within the extended range. | ||
| Calibrator | CRP T Standard (K954992) | Cfas Proteins (K012393) | The acceptance criteria for the new calibrator (Cfas Proteins) would be that its use does not adversely affect the fundamental performance characteristics (precision, accuracy, range) of the assay, implicitly demonstrated by the other performance studies. |
| Quality Control | CRP T Control (K954992), CRP T Control N (K003400) | CRP T Control N, Precinorm Protein (K012371), Precipath Protein (K012371) | Similar to the calibrator, the acceptance criteria for the new quality controls (Precinorm Protein, Precipath Protein) would be that they enable effective quality monitoring without negatively impacting assay performance. |
| Reagent Stability (On-board) | 12 weeks | 12 weeks (Note: On-board stability on cobas c 111, 5 weeks) | For most platforms, the on-board stability remains "Same," meeting the previous acceptance standard. For the cobas c 111, the acceptance criterion is 5 weeks, presumably established through specific testing for that platform. |
2. Sample Sizes and Data Provenance
-
Test Set Sample Sizes:
- Method Comparison (Modified Device vs. Predicate): n=150
- Predicate Method Comparison (old data): n=244
- Lipemia Interference: The text provides L-index values, implying samples were tested, but does not list specific sample numbers.
- Precision: Not explicitly stated for the "test set" but generally relies on replicates of samples across different levels.
-
Data Provenance: The text does not explicitly state the country of origin of the data or whether it was retrospective or prospective. Given it is a modification of an existing device by Roche Diagnostics (headquartered in Indianapolis, IN, USA, but a global company), and it involves specific instrumentation (COBAS Integra, Roche/Hitachi cobas c), it is likely prospective testing conducted in a laboratory setting, potentially at multiple sites, to validate the modifications. Without further details, specific provenance cannot be determined.
3. Number of Experts and Qualifications for Ground Truth
- This device is an in vitro diagnostic (IVD) device for quantitative determination of C-reactive protein. For such devices, "ground truth" is typically established by:
- Reference Methods: Comparison against established, well-characterized reference methods or highly accurate existing assays.
- Trueness/Accuracy to Certified Reference Materials: Using materials with assigned values traceable to international standards (e.g., IFCC/BCR/CAP reference preparation CRM 470 as mentioned for traceability).
- The concept of "number of experts" and "qualifications of experts" (like radiologists) for ground truth is generally not applicable to this type of IVD device. The ground truth is determined by the analytical performance against known standards and other validated methods, not by human interpretation of images or observations.
4. Adjudication Method
- Not applicable. Adjudication methods (e.g., 2+1, 3+1) are typically used in studies where human interpretation of data (like medical images in AI studies) requires consensus or tie-breaking. For an IVD device like C-Reactive Protein (Latex), the results are quantitative numerical values, and the comparison is statistical against reference methods or the predicate, not through expert adjudication of results.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- No. An MRMC study is not relevant for this type of IVD device. MRMC studies are used for evaluating diagnostic imaging systems or AI systems where human readers interpret cases, and the study assesses how AI assistance impacts human reader performance. This device is an automated biochemical assay, not an imaging interpretation tool.
6. Standalone Performance (Algorithm Only without Human-in-the-Loop)
- Yes, this is a standalone device. The C-Reactive Protein (Latex) assay, as an automated turbidimetric assay, operates independently to produce a quantitative result. There is no human "in the loop" during the measurement process itself, beyond loading samples and reagents and interpreting the final numerical result. The performance characteristics (precision, linearity, method comparison, interference) represent the standalone performance of the assay on the specified instruments.
7. Type of Ground Truth Used
- Reference Methods and Traceability to Certified Reference Materials: The "Method Comparison" section compares the device to a predicate device (K981897), which itself was likely validated against a reference method (the predicate's comparison was against "non-latex CRP (K951595)"). The device also states traceability to the IFCC/BCR/CAP reference preparation CRM 470 for 14 serum proteins. This indicates that the "ground truth" for calibrating and validating the assay's accuracy is tied to established international reference materials and methods.
8. Sample Size for the Training Set
- Not explicitly stated in the provided text, and likely not applicable in the typical "training set" sense for AI/machine learning. This is a chemical assay, not an AI/machine learning algorithm that requires a "training set" in that context. The development process includes formulation, optimization, and extensive validation with numerous samples, but these are generally considered part of the R&D and validation phases, rather than a distinct "training set" as understood in AI.
9. How the Ground Truth for the Training Set Was Established
- Not applicable in the typical AI sense. Since the device is a chemical assay rather than an AI model requiring a training phase, there isn't a "training set" with ground truth established in the same way. The accuracy and performance of the assay are intrinsically linked to the chemical design, reagent quality, instrument parameters, and calibration against reference materials (like CRM 470), which are developed and optimized through standard laboratory practices and analytical validation.
§ 866.5270 C-reactive protein immunological test system.
(a)
Identification. A C-reactive protein immunological test system is a device that consists of the reagents used to measure by immunochemical techniques the C-reactive protein in serum and other body fluids. Measurement of C-reactive protein aids in evaluation of the amount of injury to body tissues.(b)
Classification. Class II (performance standards).