(266 days)
The Randox RX daytona plus magnesium (Mg) test system is intended for the quantitative in vitro determination of magnesium concentration in serum, urine and lithium heparinized plasma. Magnesium measurements are used in the diagnosis and treatment of hypomagnesemia (abnormally low levels of magnesium) and hypermagnesemia (abnormally high levels of magnesium).
The Magnesium kit assay consists of a ready to use reagent solution.
The document describes the analytical performance characteristics of the Randox RX daytona plus magnesium (Mg) test system, which is a quantitative in vitro diagnostic device for measuring magnesium levels in serum, urine, and lithium heparinized plasma. The study aims to demonstrate substantial equivalence to a predicate device, the Siemens Magnesium (MG) test system (K991576).
Here's a breakdown of the requested information based on the provided text:
1. A table of acceptance criteria and the reported device performance
Please note that the document does not explicitly state predetermined acceptance criteria for all performance characteristics. Instead, it often describes the methodology and then presents the results. For some sections, like Analytical Specificity, an acceptance criterion is mentioned. I will infer or state the presented performance for others.
| Performance Characteristic | Acceptance Criteria (explicit or implicit) | Reported Device Performance |
|---|---|---|
| Precision | No explicit numerical acceptance criteria stated; inferred to be comparable to typical IVD performance for magnesium assays. The results presented should demonstrate low variability (SD, CV%). | Serum: |
| QC1 (2.36 mg/dl): Total SD 0.07, CV 2.8% | ||
| QC2 (4.36 mg/dl): Total SD 0.14, CV 3.3% | ||
| Serum Pool 1 (0.90 mg/dl): Total SD 0.04, CV 4.1% | ||
| Urine: | ||
| Urine Pool 1 (3.15 mg/dl): Total SD 0.18, CV 5.8% | ||
| LIN (21.10 mg/dl): Total SD 1.23, CV 5.8% | ||
| Linearity/Reportable Range | Deviation from linearity less than 5%. The reportable range should encompass clinically relevant magnesium levels. | Serum: Linear Regression Y = 0.96x + 0.08, r = 0.999. Reportable range: 0.74 – 4.95 mg/dl. |
| Urine: Linear Regression Y = 0.97x + 0.32, r = 0.998. Reportable range: 1.01 – 23.82 mg/dl. | ||
| Detection Limit | Limit of Quantitation (LoQ) with a %CV of ≤20%. LoD and LoB also determined. | Serum: LoB 0.28 mg/dl, LoD 0.39 mg/dl, LoQ 0.55 mg/dl (with %CV ≤20%). |
| Urine: LoB 0.44 mg/dl, LoD 0.68 mg/dl, LoQ 0.95 mg/dl (with %CV ≤20%). | ||
| Analytical Specificity / Interference | % of Control ± 10% for tested interferents. | Serum: No significant interference for Hemoglobin (up to 1000mg/dl), Total Bilirubin (up to 60mg/dl), Conjugate Bilirubin (up to 60mg/dl), Triglycerides (up to 2000mg/dl), Intralipid® (up to 500mg/dl), Ascorbic Acid (up to 6mg/dl) at Mg concentrations of 3.89 mg/dl and 6.32 mg/dl. |
| Urine: No significant interference for various analytes at 4.87mg/dl and 24.33mg/dl Mg concentrations (e.g., Direct Bilirubin 60mg/dl, Glucose 2000mg/dl, Sodium Chloride 4000mg/dl). | ||
| Method Comparison with Predicate Device | Correlation coefficient (r) ideally close to 1.0, and regression equation (Y=mx+c) with slope (m) close to 1.0 and y-intercept (c) close to 0.0, indicating strong agreement with the predicate device. | Serum: Y = 0.994x + 0.050, r = 0.992. (Compared to Siemens Magnesium (MG) on Advia 1800). |
| Urine: Y = 0.990x + 0.067, r = 0.999. (Compared to Siemens Magnesium (MG) on Advia 1800). | ||
| Matrix Comparison | Correlation coefficient (r) close to 1.0, and regression equation (Y=mx+c) for serum vs. lithium heparin plasma demonstrating equivalent results. | Y = 0.96x + 0.09, r = 0.992. (Serum vs. Lithium Heparin Plasma). |
The studies described in the document, demonstrating good precision (low CVs), linearity over the stated ranges, low detection limits, minimal interference from common analytes, and strong correlation with the predicate device, collectively prove that the device meets the implicit acceptance criteria for analytical performance of a magnesium test system.
2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
-
Precision Test Set: Not explicitly stated as "test set" in the context of ground truth, but for the precision study:
- Serum: 5 different levels of unaltered human serum samples, spiked or diluted. Each level run in 80 replicates (2 replicates per run for 20 non-consecutive days, across 2 systems). So, a total of 5 levels * 80 replicates = 400 measurements for serum samples, plus control samples.
- Urine: 3 levels of human urine supplemented with magnesium chloride, plus one "LIN" sample (normal urine pool spiked). Each level run in 80 replicates. So, a total of 4 levels * 80 replicates = 320 measurements for urine samples, plus control samples.
- Data Provenance: Not explicitly stated (e.g., country of origin, retrospective or prospective). The use of "unaltered human serum samples" and "human urine supplemented with magnesium chloride" suggests real human samples, and the study design implies a prospective collection for the purpose of the study.
-
Linearity Test Set:
- Serum & Urine: 11 levels of samples, created by mixing low and high serum/urine pools. Each level run in 5 replicates. This means 11 levels * 5 replicates = 55 measurements per matrix (serum/urine).
- Data Provenance: Not explicitly stated. The samples were prepared from "low and high serum pools," indicating human samples were used as a base.
-
Detection Limit Test Set:
- Serum & Urine: 4 low-level samples for LoD/LoB/LoQ. Based on 240 determinations.
- Data Provenance: Not explicitly stated.
-
Analytical Specificity / Interference Test Set:
- Serum & Urine: Not specific sample sizes per interferent listed, but analytes tested at specific magnesium concentrations (e.g., 3.89 mg/dl and 6.32 mg/dl for serum; 4.87mg/dl and 24.33mg/dl for urine). Interferent levels tested are specified.
- Data Provenance: Not explicitly stated. The samples were likely prepared in-house by spiking interferents into human control matrices.
-
Method Comparison Test Set:
- Serum: 108 serum patient samples.
- Urine: 108 urine patient samples.
- Data Provenance: Not explicitly stated (e.g., country of origin, retrospective or prospective). These are "patient samples," which typically implies retrospective or prospectively collected clinical samples.
-
Matrix Comparison Test Set:
- Serum vs. Lithium Heparin Plasma: A minimum of 42 matched patient sample pairs (serum and lithium heparin plasma).
- Data Provenance: Not explicitly stated. These are "patient samples," implying clinical samples.
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 device is an in vitro diagnostic (IVD) for quantitative measurement of magnesium, not an imaging device or a device requiring human interpretation for "ground truth" in the typical sense of expert consensus. The ground truth for such devices is established through reference methods, traceability to certified reference materials, and the inherent analytical measurement of the analyte.
- Reference Methods: The predicate device itself (Siemens Magnesium (MG)) serves as the "reference" for method comparison.
- Traceability: The Randox Calibration Serum Level 3 is stated to be traceable to Magnesium reference material NIST 909b. This NIST standard is the ultimate "ground truth" for magnesium concentration.
- No human experts are mentioned or typically involved in establishing the "ground truth" for the concentration values in these types of analytical studies. The "ground truth" is analytical, not interpretive.
4. 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 in studies where human readers provide interpretations and discrepancies need to be resolved. This document describes analytical performance studies of a laboratory diagnostic assay, where quantitative results are compared to known concentrations or a predicate device. There is no human interpretation involved in generating the "ground truth" for the concentrations themselves, nor in interpreting the results of the device in a way that would require 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
Not applicable. This is an in vitro diagnostic (IVD) device for quantitative measurement of magnesium, not an AI-powered imaging device or a device requiring human readers/interpreters. Therefore, no MRMC study was conducted, and there's no concept of human readers improving with or without AI assistance for this type of device.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done
This is an analytical instrument and reagent system. By its nature, its performance is "standalone" in generating a quantitative result. The device (Randox RX daytona plus system with Randox Magnesium reagents) processes samples and provides a numerical magnesium concentration. There is no "human-in-the-loop" in the sense of modifying or assisting the algorithmic output of the concentration measurement. The operator's role is to load samples and reagents and initiate the automated analysis, and then review the instrument's quantitative output.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
The ground truth for the analytical performance studies is established by:
- Reference materials/standards: Traceability to NIST 909b for calibrators, and the use of control materials with known concentrations.
- Known sample preparations: For linearity and detection limit studies, samples are often prepared by spiking or diluting to known target concentrations.
- Predicate device: For method comparison, the results obtained from the new device are compared to results obtained from a legally marketed predicate device (Siemens Magnesium (MG) on Advia 1800), which itself is established as accurate.
There is no expert consensus, pathology, or outcomes data used to establish the "ground truth" for magnesium concentration values in these studies.
8. The sample size for the training set
The provided document describes studies for demonstrating analytical performance and substantial equivalence to a predicate device. These are validation studies, not machine learning or AI development studies that typically involve "training sets." Therefore, the concept of a training set as understood in AI/ML is not applicable here, and no training set sample size is mentioned.
9. How the ground truth for the training set was established
As there is no "training set" in the context of AI/ML, this question is not applicable. The ground truth for the validation of the device (as discussed in point 7) is established through analytical traceability, known preparations, and comparison to an established predicate device.
§ 862.1495 Magnesium test system.
(a)
Identification. A magnesium test system is a device intended to measure magnesium levels in serum and plasma. Magnesium measurements are used in the diagnosis and treatment of hypomagnesemia (abnormally low plasma levels of magnesium) and hypermagnesemia (abnormally high plasma levels of magnesium).(b)
Classification. Class I.