(367 days)
iFlash-HCG is a paramagnetic particle chemiluminescent immunoassay (CLIA) for quantitative detection of the intact human chorionic gonadotropin (hCG) molecule and the hCG ß-subunit (ß-hCG) in human serum and plasma using the automated Chemiluminescence Immunoassay Analyzer (Model: iFlash-HCG assay is to be used by laboratory professionals as an aid in early detection of pregnancy together with other clinical methods.
Chemiluminescence Immunoassay Analyzer (Model: iFlash 3000-C) is a fully-automated, chemiluminescence immunoassay analyzer intended for quantitative or qualitative determination of analytes in human body fluids taken from clinical settings. It is used together with its supporting chemiluminescence immunoassay reagence Immunoassay Analyzer (Model: iFlash 3000-C) is intended for use in clinical laboratories.
iFlash-HCG that includes testing reagents and three levels of calibrators is based on chemiluminescence immunoassay. HCG and hCG ß-subunit (ß-hCG) in the sample reacts with anti-HCG antibody coated paramagnetic microparticles and acridinium-labeled anti-HCG antibody conjugate to form a sandwich complex, after chemiluminescent reaction, HCG amount in the sample is derived from RLUs (relative light units) using a calibration curve. iFlash-HCG is intended to be used on Chemiluminescence Immunoassay Analyzer (Model: iFlash 3000-C).
Chemiluminescence Immunoassay Analyzer (Model: iFlash 3000-C) is a fully-automated, chemiluminescence immunoassay analyzer intended for quantitative determination of analytes in human body fluids taken from clinical settings. It is used together with its supporting chemiluminescence immunoassay reagents. The Chemiluminescence Immunoassay Analyzer (Model: iFlash 3000-C) is intended for use in clinical laboratories.
The provided text describes the performance of the iFlash-HCG and Chemiluminescence Immunoassay Analyzer (Model: iFlash 3000-C) for the quantitative detection of human chorionic gonadotropin (hCG). However, it does not detail acceptance criteria in a structured table or specifically describe a "study that proves the device meets the acceptance criteria" in terms of clinical performance against defined benchmarks with human experts, as would be typical for an AI/ML-based diagnostic device where acceptance criteria often relate to sensitivity, specificity, and agreement with ground truth.
Instead, the document focuses on detailed non-clinical performance studies demonstrating the analytical characteristics and substantial equivalence to a predicate device. Many of the listed studies (e.g., precision, detection capability, linearity, interference, analytical specificity, method comparison, stability, trueness, sample dilution, reference interval, carryover) are standard for in vitro diagnostic (IVD) devices.
Given the input, I will interpret "acceptance criteria" as the performance specifications demonstrated by the non-clinical studies and "study that proves the device meets the acceptance criteria" as the results of these non-clinical studies. I will also clarify that this is not an AI/ML device per se, so the typical AI/ML study components (experts for ground truth, MRMC, standalone algorithm performance) are not applicable in their traditional sense.
Here's the information extracted and structured based on your request, with caveats where the information is not present or not applicable to an IVD device of this type:
1. Table of Acceptance Criteria and Reported Device Performance
As the document does not provide a pre-defined table of "acceptance criteria" in the sense of pass/fail thresholds for clinical performance but rather lists the results of various analytical performance studies, I will present the key performance parameters and their achieved values. The "acceptance criteria" here are implicitly met if the reported performance is deemed suitable for the intended use and demonstrates substantial equivalence to predicate devices.
| Performance Parameter | Acceptance Criteria (Implicit / Demonstrated Performance) | Reported Device Performance (as presented) |
|---|---|---|
| Precision | Demonstrated according to CLSI EP05-A3 standard | Repeatability and Reproducibility SD and CV% calculated per EP05-A3. (Specific values not provided in summary, but stated to demonstrate fulfillment). |
| Detection Limit (LoB) | Determined according to CLSI EP17-A2 | LoB = 0.10 mIU/mL |
| Detection Limit (LoD) | Determined according to CLSI EP17-A2 | LoD = 0.20 mIU/mL |
| Detection Limit (LoQ) | Determined according to CLSI EP17-A2 | LoQ = 0.50 mIU/mL (Total error limit ≤30%) |
| Linearity Range | Demonstrated according to CLSI EP06 2nd Edition | 0.50 - 10000.00 mIU/mL (predefined allowable deviation ±15%) |
| Hook Effect | No Hook effect observed within concentration | No HOOK effect observed within HCG/ß-HCG concentration of 1,250,000 mIU/mL. |
| Interference (Endogenous) | Not susceptible to interference at specified levels | Not susceptible at levels: Bilirubin (conj. ≤40mg/dL, unconj. ≤40mg/dL), Hemoglobin (≤1000mg/dL), Triglyceride (≤3000 mg/dL), Serum total protein (≤10 g/dL), Rheumatoid factors (2000 IU/mL), HAMA (600 ng/mL), ANA (500 AU/mL). |
| Interference (Exogenous/Drug) | Not susceptible to interference at specified levels | Not susceptible at specified levels for 17 common drugs. |
| Analytical Specificity (Cross-reactivity) | Not susceptible to interference from specified cross-reactants | Not susceptible to LH (500 mIU/mL), FSH (200 mIU/mL), TSH (10000 mIU/mL). |
| Specimen Types Comparison | Good agreement between serum and plasma samples | Paired serum and plasma samples in good agreement (Passing-Bablok Regression). |
| Method Comparison | Good consistency with predicate device | Y=0.986X-0.047; correlation coefficient T= 0.998 (110 serum samples covering 0.531-9717 mIU/mL). |
| Stability | Data supports claims in user manual | Stability data supports claims. |
| Trueness | Relative deviation within ±10.0% | Achieved for 25.00, 200.00, 4000.00 mIU/mL samples against WHO standard. |
| Sample Dilution Recovery | Relative deviation within ±10% | Supports dilution of samples with HCG concentrations above 10000 mIU/mL with max 1:100 dilution. |
| Reference Interval | Established through study | Non-pregnant premenopausal women (18-50): 95th percentile 0.6 mIU/mL (N=130); Postmenopausal women (≥50): 95th percentile 5.4 mIU/mL (N=125). |
| Carryover Effect | No carryover effect observed | No carryover effect observed with high HCG (≥1,000,000 mIU/mL) followed by low HCG (≤5 mIU/mL). |
2. Sample Size Used for the Test Set and Data Provenance
-
Test Set (for performance studies):
- Precision: 9 levels of female serum samples and 2 levels of controls (total samples tested over 20 days: 9 samples * 2 runs * 2 replicates * 20 days * 3 lots * 3 analyzers = substantial, specific number of unique patients/samples not given).
- Detection Limit (LoB): 5 analyte-free samples * 4 times/day * 3 days * 3 reagent lots (60 results per reagent lot).
- Detection Limit (LoD): 5 low-concentration samples * 4 times/day * 3 days * 3 reagent lots (60 test results per reagent lot).
- Detection Limit (LoQ): 5 samples near LoD * 4 times/day * 3 days * 3 reagent lots (60 test results per reagent lot).
- Linearity: 11 different concentration levels of samples for each of 3 linearity intervals tested on 3 reagent lots.
- Hook Effect: 3 high concentration samples and their serial dilutions.
- Interference Study: Not specified, but likely involved multiple samples spiked with interferents.
- Analytical Specificity: Not specified, but likely involved multiple samples spiked with cross-reactants.
- Specimen Types Study: 97 female serum samples compared with plasma samples.
- Method Comparison: 110 serum samples.
- Trueness Study: Samples formulated from WHO International Standard (3 concentration levels).
- Sample Dilution Fold Study: Samples prepared with HCG positive material (3 theoretical concentrations).
- Reference Interval Study: Non-pregnant premenopausal women (N=130), Postmenopausal women (N=125).
- Carryover Study: Test samples with high HCG in triplicate followed by low HCG in triplicate, for five runs.
-
Data Provenance: The document does not explicitly state the country of origin for the clinical samples used in these studies. It does indicate that the submitter is based in Shenzhen, Guangdong, China. The studies are described as non-clinical performance studies, often implying controlled laboratory conditions rather than broad population-based data collection. All studies appear to be prospective as they are specifically conducted to evaluate device performance.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications of those Experts
This is an IVD device for quantitative measurement of hCG, not an AI/ML medical image analysis or similar diagnostic device that typically relies on human expert interpretation for "ground truth." The ground truth for this device's performance is established by reference methods, certified reference materials (like WHO International Standard 5th WHO IS Chorionic Gonadotrophin 07/364), and comparison with legally marketed predicate devices. Therefore, the concept of "experts" establishing conventional ground truth as applied to AI/ML clinical studies is not applicable here.
4. Adjudication Method for the Test Set
As there are no human experts classifying or interpreting data for "ground truth" in the AI/ML sense, there is no adjudication method described or applicable. The determination of results is based on the chemical reaction and optical detection by the automated analyzer, with analytical results compared to established reference values or predicate device results.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
An MRMC study is relevant for diagnostic devices (often imaging-based or AI-assisted) where human reader performance is a key outcome. This document describes an automated IVD assay, not an imaging device or an AI assistance tool for human readers. Therefore, no MRMC comparative effectiveness study was done, and the concept of human readers improving with AI assistance is not applicable.
6. Standalone (i.e., algorithm only without human-in-the-loop performance) Study
This device is a standalone automated analyzer ("algorithm only" in the sense of the instrument performing the test independently). The entire suite of non-clinical performance studies (precision, detection limits, linearity, interference, method comparison, etc.) constitutes the demonstration of its standalone analytical performance. It does not require human-in-the-loop for its operation or result generation in the same way an AI diagnostic algorithm might.
7. The Type of Ground Truth Used
The "ground truth" for the analytical validation of this IVD device is primarily based on:
- Reference materials/standards: Notably, the WHO International Standard 5th WHO IS Chorionic Gonadotrophin 07/364 is used for trueness studies.
- Spiked samples: Known concentrations of analytes or interferents are added to samples to confirm recovery and specificity.
- Comparison to predicate devices: The results from the iFlash-HCG are compared against those from the cleared predicate devices (Elecsys HCG+β reagent and Cobas e 801 analyzer) to demonstrate substantial equivalence.
- Consensus methods/protocols: CLSI guidelines (EP05-A3, EP17-A2, EP06 2nd Edition, EP07-A3, EP37 1st Edition, EP09c 3rd Edition, EP25-A, EP34 1st Edition, EP28-A3c, H26-A2) provide the "ground truth" for how studies should be designed and how performance metrics should be calculated and interpreted.
8. The Sample Size for the Training Set
This document describes a conventional IVD assay and analyzer, not an AI/ML system that undergoes a distinct "training" phase with a large dataset. Therefore, the concept of a "training set" for an AI/ML algorithm is not applicable in this context. The methodology relies on established chemical and physical principles, not machine learning from data.
9. How the Ground Truth for the Training Set Was Established
As there is no "training set" in the AI/ML sense, this question is not applicable. The device's performance is inherently determined by its design specifications, reagents, and analytical principles, validated through the non-clinical studies.
§ 862.1155 Human chorionic gonadotropin (HCG) test system.
(a)
Human chorionic gonadotropin (HCG) test system intended for the early detection of pregnancy —(1)Identification. A human chorionic gonadotropin (HCG) test system is a device intended for the early detection of pregnancy is intended to measure HCG, a placental hormone, in plasma or urine.(2)
Classification. Class II.(b)
Human chorionic gonadotropin (HCG) test system intended for any uses other than early detection of pregnancy —(1)Identification. A human chorionic goadotropin (HCG) test system is a device intended for any uses other than early detection of pregnancy (such as an aid in the diagnosis, prognosis, and management of treatment of persons with certain tumors or carcinomas) is intended to measure HCG, a placental hormone, in plasma or urine.(2)
Classification. Class III.(3)
Date PMA or notice of completion of a PDP is required. As of the enactment date of the amendments, May 28, 1976, an approval under section 515 of the act is required before the device described in paragraph (b)(1) may be commercially distributed. See § 862.3.