(34 days)
The Total Bilirubin assay is used for the quantitation of total bilirubin in human serum or plasma of adults and neonates on the ARCHITECT c8000 System.
Measurement of total bilirubin, an organic compound formed during the normal destruction of red blood cells, is used in the diagnosis and treatment of liver, hematological and metabolic disorders, including hepatitis and gall bladder block. A bilirubin (total and unbound) in the neonate test system is a device intended to measure the levels of bilirubin (total and unbound) in the blood (serum) of newborn infants to aid in indicating the risk of bilirubin encephalopathy (kernicterus).
The Total Bilirubin reagent kit contains Reagent 1 and Reagent 2. Reagent 1 contains Surfactants and HCl. Reagent 2 contains 2, 4-dichloroaniline, HCl, Sodium nitrite, and Surfactant. The principles of the procedure involve total (conjugated and unconjugated) bilirubin coupling with a diazo reagent in the presence of a surfactant to form azobilirubin. The increase in absorbance at 548 nm due to azobilirubin is directly proportional to the total bilirubin concentration. The detection of the analyte is end-point colorimetric.
Here's an analysis of the provided text regarding the acceptance criteria and study for the Total Bilirubin device:
1. Table of Acceptance Criteria and Reported Device Performance:
The document doesn't explicitly state "acceptance criteria" for each test in a formal table, but rather describes the methodology and then reports the results. I will infer the acceptance criteria from the statements provided about what is considered acceptable or how performance supports the claims.
| Performance Characteristic | Acceptance Criteria (Inferred from text) | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | Support the lower end of the measuring interval (0.3 mg/dL). | The LoQ of the Total Bilirubin assay is 0.174 mg/dL. The LoQ supports the lower end of the measuring interval of 0.3 mg/dL. LoD is 0.174 mg/dL and LoB is 0.102 mg/dL. |
| Linearity | Demonstrate linearity across the measuring interval of 0.3 to 25.0 mg/dL. | The Total Bilirubin assay was demonstrated to be linear across the measuring interval of 0.3 to 25.0 mg/dL. |
| Precision (Within-Laboratory/Total) | Not explicitly stated as a numerical criterion in %CV or SD, but the study was conducted according to CLSI EP5-A2, implying adherence to industry-standard precision expectations. | |
| Sample 1 (0.87 mg/dL): Within Run SD 0.010 mg/dL (1.2%CV), Total SD 0.018 mg/dL (2.1%CV). | ||
| Sample 2 (4.53 mg/dL): Within Run SD 0.034 mg/dL (0.8%CV), Total SD 0.057 mg/dL (1.3%CV). | ||
| Sample 3 (6.42 mg/dL): Within Run SD 0.041 mg/dL (0.6%CV), Total SD 0.066 mg/dL (1.0%CV). | ||
| Sample 4 (16.85 mg/dL): Within Run SD 0.115 mg/dL (0.7%CV), Total SD 0.204 mg/dL (1.2%CV). | ||
| Interference | Bias of > 10%, or > 0.2 mg/dL (> 3.4 umol/L) for bilirubin concentrations ≤ 2.0 mg/dL is considered significant interference. | Hemoglobin (2000 mg/dL): At 1.0 mg/dL, -9.8% bias. At 13.4 mg/dL, -3.0% bias. |
| Intralipid (1000 mg/dL): At 1.0 mg/dL, -0.6% bias. At 13.4 mg/dL, -0.6% bias. | ||
| Indican (0.125 mmol/L): At 0.9 mg/dL, 26.7% bias. At 14.7 mg/dL, 3.2% bias. | ||
| Conclusion: Indican, at concentrations > 0.125 mmol/L, interferes with the assay in samples with bilirubin levels at 1.2 mg/dL. | ||
| Method Comparison (Predicate) | Correlation coefficient and regression line (slope close to 1, intercept close to 0) indicating substantial equivalence to the predicate. | Adult Population (N=118): Slope 0.99, Intercept -0.09, Correlation Coefficient 0.9994. Range: 0.3 to 24.8 mg/dL. |
| Neonate Population (N=54): Slope 0.96, Intercept 0.01, Correlation Coefficient 0.9982. Range: 0.3 to 24.3 mg/dL. | ||
| Reference Range | Not explicitly stated as a numerical criterion for the study, but the observed range determined should be clinically acceptable. | The reference range was determined to be 0.3 mg/dL to 1.2 mg/dL. 4 of 40 samples (10%) were outside this range, all being 25 mg/dL, either System Automated Dilution or Manual Dilution protocol can be performed. |
| Specimen Tube Type | Deming linear regression analysis (slope close to 1, intercept close to 0, high correlation) demonstrating equivalence to the control tube type (serum plastic). | Serum (glass) (N=41): Slope 0.96, Intercept 0.06, Correlation Coefficient 0.9990. |
| SST (N=40): Slope 1.00, Intercept 0.00, Correlation Coefficient 0.9996. | ||
| EDTA (N=39): Slope 1.00, Intercept -0.01, Correlation Coefficient 0.9990. | ||
| Lithium Heparin (N=40): Slope 1.01, Intercept 0.02, Correlation Coefficient 0.9994. | ||
| Lithium Heparin Plasma Separator Tube (PST) (N=40): Slope 0.96, Intercept 0.05, Correlation Coefficient 0.9992. | ||
| Sodium Heparin (N=39): Slope 0.98, Intercept 0.03, Correlation Coefficient 0.9996. | ||
| Conclusion: All tested tube types are acceptable, except those containing sodium fluoride/potassium oxalate due to hemolysis potential. |
2. Sample Sizes Used for the Test Set and Data Provenance:
- Limit of Quantitation (LOQ), Limit of Detection (LOD), Limit of Blank (LOB):
- Zero-analyte samples: 4 samples. Tested in a minimum of 5 replicates on 5 separate runs.
- Low-analyte samples: Minimum of 2 samples gravimetrically prepared at 8 target concentrations. Tested in a minimum of 10 replicates on 5 separate runs.
- Data Provenance: Not explicitly stated (e.g., country of origin) but "human serum albumin" and "unconjugated bilirubin" are used. The study is prospective in nature, as samples are prepared for the purpose of the test.
- Linearity: 12 levels per pool in each of three combined bilirubin pools. Tested in a minimum of 4 replicates.
- Data Provenance: "combined bilirubin pools" from "conjugated bilirubin stock" and "unconjugated bilirubin stock", "serum". Prospective.
- Within-Laboratory Precision: 4 control materials (Bio-Rad serum based). Tested in a minimum of 2 replicates, 2 times per day for 20 days.
- Data Provenance: Commercial control materials (Bio-Rad Lyphochek Unassayed Chemistry Control, Bio-Rad Liquichek Pediatric Control). Prospective.
- Interference: Not specified.
- Data Provenance: Not specified, but uses "bilirubin concnetrations". Prospective.
- Method Comparison:
- Adult patient specimens: 124. 4 spiked.
- Neonatal patient specimens: 64. 4 spiked.
- Data Provenance: Patient specimens. Not specified if retrospective or prospective or country of origin, but generally method comparison studies use collected patient samples.
- Reference Range: 40 adult patient serum samples.
- Data Provenance: Fresh, adult patient serum samples from a clinically healthy population. Stored at 2-8°C, protected from light. Assumed to be prospective as samples were collected for the study.
- Automated Dilution Protocol versus Manual Dilution Procedure: Not specified (samples were pooled to create desired concentrations).
- Data Provenance: "Fresh serum specimens" obtained and pooled. Prospective.
- Specimen Tube Type (Matrix Equivalence): Minimum of 40 samples from adult subjects for each tube type. (Serum glass N=41, SST N=40, EDTA N=39, Lithium Heparin N=40, PST N=40, Sodium Heparin N=39).
- Data Provenance: "Fresh or frozen sample sets" from subjects. Assumed to be prospective as these are explicitly collected to test tube types.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts:
- This device is an in vitro diagnostic (IVD) for quantitative measurement. The "ground truth" for such devices is typically established through analytical methods, reference materials, gravimetric preparation, or comparison to a gold standard reference method/device, rather than expert human interpretation.
- Therefore, the concept of "experts used to establish the ground truth" as it applies to image-based AI or clinical diagnostic interpretation by physicians is not applicable here. The ground truth is the chemical concentration of bilirubin.
4. Adjudication Method for the Test Set:
- Again, as this is a quantitative chemical measurement, adjudication methods for expert interpretation (like 2+1, 3+1) are not applicable. The "adjudication" is inherent in the analytical process (e.g., repeating measurements, using certified reference materials, performing statistical analysis of replicates).
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:
- This is an in vitro diagnostic (IVD) device. MRMC studies, which assess human reader performance with and without AI assistance, are typically conducted for AI-powered medical image analysis or clinical decision support systems.
- Therefore, an MRMC comparative effectiveness study is not applicable to this type of device. There are no "human readers" interpreting an output in the same way a radiologist reads an image.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done:
- Yes, this entire submission is a standalone performance evaluation of the assay (the "algorithm only" in a broader sense of the measurement procedure). The device is designed to quantitatively measure bilirubin, and the tests (LOQ, linearity, precision, interference, method comparison) evaluate its analytical performance without direct human interpretive intervention beyond running the instrument and analyzing the data.
7. The Type of Ground Truth Used:
- For LOQ, LOD, LOB: Gravimetrically prepared samples (human serum albumin and unconjugated bilirubin) provide the known "ground truth" concentrations.
- For Linearity: Combined bilirubin pools with known proportional compositions.
- For Precision: Commercial control materials with established (though perhaps unassayed) target ranges, tested repeatedly.
- For Method Comparison: The predicate device's measurements are used as the comparative "ground truth" or reference, as the goal is to show substantial equivalence.
- For Reference Range: Clinical health status of the adult population samples.
- For Automated Dilution: Known target concentrations and manual dilution results.
- For Specimen Tube Type: The control tube type (serum plastic tube) serves as the reference for comparison.
8. The Sample Size for the Training Set:
- This document describes the analytical validation of a re-agent kit for an existing instrument (ARCHITECT c8000 System). It's not an AI model that undergoes "training" in the conventional sense.
- Therefore, the concept of a "training set" for an AI algorithm is not applicable. The development of the reagent and its underlying chemical principles involved R&D and optimization, but not machine learning training on a dataset.
9. How the Ground Truth for the Training Set Was Established:
- As explained above, there is no "training set" in the context of an AI algorithm described here. The "ground truth" used during the development of the assay would have been based on established clinical chemistry principles, reference methods, and gravimetric preparations to ensure accurate concentration measurements during formulation and optimization.
§ 862.1110 Bilirubin (total or direct) test system.
(a)
Identification. A bilirubin (total or direct) test system is a device intended to measure the levels of bilirubin (total or direct) in plasma or serum. Measurements of the levels of bilirubin, an organic compound formed during the normal and abnormal distruction of red blood cells, if used in the diagnosis and treatment of liver, hemolytic hematological, and metabolic disorders, including hepatitis and gall bladder block.(b)
Classification. Class II.