(157 days)
The CRE2 method is an in vitro diagnostic test for the quantitative measurement of creatinine in human serum, plasma, and urine on the Dimension® clinical chemistry system. Creatinine measurements are used in the diagnosis and treatment of certain renal disease, in monitoring renal dialysis, and as a calculation basis for measuring other urine analytes.
The CRE2 method uses a modified kinetic Jaffe technique. In the presence of a strong base such as sodium hydroxide, picrate reacts with creatinine to form a red chromophore. The rate of increasing absorbance at 510 nm due to the formation of this chromophore is directly proportional to the creatinine concentration in the sample and is measured using a bichromatic (510, 600nm) rate technique. Bilirubin is oxidized by potassium ferricyanide to prevent interference.
The provided text describes a 510(k) premarket notification for a new in vitro diagnostic device, the Dimension® Creatinine (CRE2) Flex® reagent cartridge. This document is a summary of the safety and effectiveness information, comparing the new device to a predicate device already on the market.
Here's an analysis of the acceptance criteria and the study that proves the device meets them:
1. A table of acceptance criteria and the reported device performance
The document does not explicitly present a table of acceptance criteria for performance parameters in a pass/fail format. Instead, it provides the results of various performance studies and implies that these results demonstrate substantial equivalence to the predicate device and/or meet established industry guidelines (CLSI). The acceptance criteria are implicitly defined by what is considered "substantially equivalent" or within acceptable limits as per CLSI guidelines.
Below is a table summarizing the reported device performance, categorized by the studies conducted. It's important to note that specific numerical acceptance criteria (e.g., "slope must be between 0.98 and 1.02") are not explicitly stated, but the reported values passed the FDA's substantial equivalence review.
| Performance Characteristic | Reported Device Performance (CRE2 Flex®) | Implied Acceptance Criteria / Context |
|---|---|---|
| Method Comparison (vs. Predicate CREA Assay) | Agreement with predicate device (CREA). | |
| Serum Sample (n=191) | Range: 0.4 - 19.8 mg/dL | Good agreement. |
| Slope: 1.00 | Ideal slope is 1.0. | |
| Intercept: -0.08 mg/dL | Ideal intercept is 0. | |
| Correlation Coefficient (r): 0.999 | Close to 1.0 indicates strong correlation. | |
| Urine Sample (n=113) | Range: 13.5 - 372.7 mg/dL | Good agreement. |
| Slope: 1.04 | Close to 1.0. | |
| Intercept: -3.58 mg/dL | Close to 0. | |
| Correlation Coefficient (r): 0.996 | Close to 1.0 indicates strong correlation. | |
| Method Comparison (vs. IDMS Reference Method) | Agreement with a recognized gold standard. | |
| Serum Sample (n=48) | Range: 0.18 - 6.32 mg/dL | Good agreement. |
| Slope: 1.04 | Close to 1.0. | |
| Intercept: 0.02 mg/dL | Close to 0. | |
| Correlation Coefficient (r): 0.997 | Close to 1.0 indicates strong correlation. | |
| Serum Plasma Equivalency | Agreement between serum and plasma samples. | |
| Lithium Heparin Plasma (n=56) | Range: 0.50 - 17.35 | - |
| Slope: 1.05 | Close to 1.0. | |
| Intercept: -0.02 | Close to 0. | |
| Correlation Coefficient (r): 0.998 | Close to 1.0 indicates strong correlation. | |
| Precision (Repeatability & Within-Lab CV) | Within acceptable limits for clinical chemistry assays. | |
| Serum Pool 1 (Mean 1.32 mg/dL) | Repeatability SD: 0.04, %CV: 3.0 | Low %CV indicates high precision. |
| Within-Lab SD: 0.04, %CV: 3.2 | Low %CV indicates high precision. | |
| Serum Pool 2 (Mean 15.79 mg/dL) | Repeatability SD: 0.19, %CV: 1.2 | Low %CV. |
| Within-Lab SD: 0.19, %CV: 1.2 | Low %CV. | |
| BioRad Multiqual Level 1 (Mean 0.71 mg/dL) | Repeatability SD: 0.03, %CV: 4.7 | Low %CV. |
| Within-Lab SD: 0.04, %CV: 5.1 | Low %CV. | |
| ... (other samples shown in table, similar low CVs) | ||
| Limit of Blank (LoB) | Meet or exceed claimed LoB. | |
| Serum LoB | 0.05 mg/dL (consistent with claim) | Claimed LoB: 0.05 mg/dL |
| Urine LoB | 0.87 mg/dL (consistent with claim) | Claimed LoB: 1.0 mg/dL |
| Limit of Detection (LoD) | Meet or exceed claimed LoD. | |
| Serum LoD | 0.08 mg/dL (consistent with claim) | Claimed LoD: 0.1 mg/dL |
| Urine LoD | 1.51 mg/dL (consistent with claim) | Claimed LoD: 2.0 mg/dL |
| Limit of Quantitation (LoQ) | Defined by allowable total error. | |
| Serum LoQ | 0.15 mg/dL (based on 0.15 mg/dL allowable total error) | - |
| Urine LoQ | 5.00 mg/dL (based on 3.00 mg/dL allowable total error) | - |
| Linearity (Measurement Range) | Demonstrate linearity across the claimed measuring range. | |
| Serum (0.15* - 22.00 mg/dL) | Slope: 1.01 | Close to 1.0. |
| Intercept: -0.01 | Close to 0. | |
| Correlation Coefficient (r): 1.0 | Indicates perfect linearity. | |
| Urine (3.00* - 425.34 mg/dL) | Slope: 0.998 | Close to 1.0. |
| Intercept: -0.41 | Close to 0. | |
| Correlation Coefficient (r): 1.0 | Indicates perfect linearity. | |
| Analytical Specificity (Interference) | Interference less than 10% bias, or disclosed. | |
| Non-interfering Substances (e.g., Acetaminophen, Caffeine, Glucose, etc.) | % Diff typically between -2% to 6% | Bias 10% (e.g., Acetone up to 11% bias) |
| HIL Interference (Hemolysis, Icterus, Lipemia) | Some substances show >10% bias (e.g., Hemoglobin at 1000 mg/dL, Bilirubin (conj) at 40 mg/dL, Bilirubin (unconj) at 20 mg/dL, Intralipid 20% at 1500 mg/dL) | Identified as interfering; implications for use would be in labeling. The predicate's stated interferences (e.g., Hemoglobin at 1000 mg/dL, Bilirubin (unconjugated) at 5 mg/dL, Lipemia (Intralipid) at 200 mg/dL) are less tolerant than the new device in some cases for Bilirubin and Lipemia. |
2. Sample sizes used for the test set and the data provenance
- Method Comparison (vs. Predicate CREA Assay):
- Serum: 191 remnant de-identified patient samples.
- Urine: 113 patient samples.
- Provenance: Remnant de-identified samples. "No patient history information was obtained on these samples." The studies were conducted internally by Siemens Healthcare Diagnostic Inc. R&D organization personnel, implying a US-based setting. The data is retrospective as it used "remnant de-identified serum/urine samples."
- Method Comparison (vs. IDMS Reference Method):
- Serum: 48 remnant de-identified patient samples.
- Provenance: Remnant de-identified samples. "No patient history information was obtained on these samples." Conducted internally by Siemens Healthcare Diagnostic Inc. R&D. Retrospective.
- Serum Plasma Equivalency:
- Samples: 56 matched serum and lithium heparin plasma samples.
- Provenance: "All samples in the study were fresh and never frozen." The eight spiked sample sets were prepared. Conducted internally by Siemens Healthcare Diagnostic Inc. R&D. Retrospective, likely sourced from a local population.
- Precision:
- Samples: Two serum pools, three levels of BioRad Multiqual material, two levels of BioRad Liquicheck material, and two urine pools. Testing involved 20 days with 2 separate runs and 2 test samples for each material. The specific number of individual patient samples is not given, as this study uses control materials and pools.
- Provenance: Not specified, but likely prepared internally or from commercial sources.
- Limit of Blank/Detection (LoB/LoD):
- Samples: 4 samples with no analyte (for LoB), 4 low patient serum/urine samples (for LoD). Each tested for 3 days, one run per day, across 2 reagent lots.
- Provenance: Not specified, but generally prepared internally with blank/low-level samples.
- Linearity (Measurement Range):
- Samples: 12 equally spaced serum samples and 12 urine samples.
- Provenance: Prepared by mixing high and low creatinine concentration samples.
- Analytical Specificity (Interference):
- Samples: Not specified in terms of number of patient samples, but involved "fresh sample pools containing either low or high levels of creatinine analyte in both serum pools and urine pools." These pools were spiked with various interfering substances.
- Provenance: Likely developed from internal pools and commercially available substances.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
This section is not applicable to this type of device (in vitro diagnostic for quantitative measurement of creatinine). The "ground truth" for clinical chemistry assays like creatinine is established by:
- Reference methods (e.g., IDMS, considered the "gold standard" for creatinine measurement in some contexts).
- Predicate devices (already cleared and established to perform reliably).
- Known concentrations in control materials or spiked samples.
There are no human "experts" (like radiologists interpreting images) involved in establishing ground truth for the test set in this context. The "personnel carrying out the study were laboratory technicians with training similar to personnel who would conduct the tests in a hospital laboratory setting."
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
This concept of "adjudication" is typical for studies involving human interpretation (e.g., medical imaging, pathology slides) where there might be disagreement among readers. For an automated in vitro diagnostic device, the measurement is quantitative and typically objective. Therefore, an adjudication method for the test set is not applicable. The "ground truth" is determined by instrumental readings and reference methods.
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
An MRMC study is not applicable here. This device is a fully automated in vitro diagnostic assay, not an AI-assisted diagnostic tool for human readers. Its performance is evaluated against chemical reference methods and a predicate device, not in terms of how it assists human interpretation.
6. If a standalone (i.e. algorithm only without human-in-the loop performance) was done
Yes, this entire submission and its performance data represent a standalone (algorithm only) performance evaluation. The device is an automated clinical chemistry system that measures creatinine levels directly from biological samples (serum, plasma, urine) without human intervention in the interpretation of the core measurement. The reported results (slope, intercept, correlation, precision, LoB/LoD, linearity, interference) directly reflect the device's inherent analytical performance.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)
The ground truth used for evaluating the Dimension® Creatinine (CRE2) Flex® reagent cartridge includes:
- Reference Method: The IDMS (Isotope Dilution Mass Spectrometry) method for creatinine measurement. This is considered a highly accurate and precise reference method.
- Predicate Device: The Creatinine Method for Use on the Dimension® Clinical Chemistry System (CREA, K925668). The predicate device's cleared performance serves as a benchmark for demonstrating substantial equivalence.
- Known Concentrations: For studies like precision, LoB/LoD, linearity, and interference, known concentrations of analytes in control materials, spiked samples, or blank samples serve as the ground truth.
8. The sample size for the training set
This document describes a 510(k) submission for an in vitro diagnostic reagent cartridge that uses a modified kinetic Jaffe technique (a traditional chemical assay), not a machine learning or AI-based device. Therefore, the concept of a "training set" for an algorithm is not applicable. The "training" for such a device would refer to its development and optimization based on chemical principles, rather than data-driven machine learning.
9. How the ground truth for the training set was established
As answered in point 8, the device does not employ a training set in the context of an AI/ML algorithm. Its "ground truth" for development and calibration would be rooted in established chemical and analytical principles, reference measurement procedures (like IDMS), and potentially existing characterized samples. The manufacturing process of reagents, calibrators, and the instrument's detection capabilities are based on fundamental scientific understanding rather than a data training process as understood in AI/ML.
§ 862.1225 Creatinine test system.
(a)
Identification. A creatinine test system is a device intended to measure creatinine levels in plasma and urine. Creatinine measurements are used in the diagnosis and treatment of renal diseases, in monitoring renal dialysis, and as a calculation basis for measuring other urine analytes.(b)
Classification. Class II.