Search Results

The i-STAT® Total Beta-Human Chorionic (B-hCG) test is an in vitro diagnostic test for the quantitative and qualitative determination of ß-hCG in venous whole blood or plasma samples using the i-STAT 1 Analyzer Systems. The test is intended to be used as an aid in the early detection of pregnancy and is for prescription use only.

The i-STAT® Total B-hCG Controls are used to monitor performance of the i-STAT Total β-hCC test.

The i-STAT® Total B-hCG Calibration Verification Materials are used to verify the calibration of the i-STAT Total BhCG test throughout the reportable range.

The i-STAT Total ß-hCG test uses a two-site enzyme-linked immunoassay with electrochemical detection of the resulting enzyme signal. Monoclonal antibodies specific for the ß subunit of hCG are localized within a lithographically defined capture region on the surface of a specially designed electrochemical sensor chip. A separate reference sensor of similar construction to the hCG sensor, but without antibodies specific to hCG, is located on the same sensor chip. An alkaline phosphatase-linked monoclonal antibody conjugate specific to a separate epitope on the ß subunit of the hCG molecule is present in a metered amount on the sensor chip.

The i-STAT Total ß-hCG Controls are assayed human serum which are used to monitor the performance of the i-STAT Total ß-hCG test.

The i-STAT Total ß-hCG Calibration Verification materials are assayed human serum used to verify the calibration of the i-STAT Total β-hCG test throughout the reportable range.

This document is a 510(k) premarket notification for the i-STAT® Total β-hCG Test, i-STAT® Total β-hCG Controls, and i-STAT® Total β-hCG Calibration Verification Materials. It aims to demonstrate substantial equivalence to legally marketed predicate devices.

1. Table of Acceptance Criteria and Reported Device Performance:

The document describes various performance characteristics, primarily for the i-STAT Total β-hCG test cartridge. While explicit "acceptance criteria" are not always presented as clear pass/fail thresholds in a dedicated table, the stated "design goals" and "defined as" criteria serve this purpose.

2. Sample sizes used for the test set and the data provenance:

• Precision / Reproducibility:
  • No specific sample size for "test set" is explicitly stated beyond "samples were targeted at four hCG concentrations" and testing at "three external POC sites."
  • Provenance: Not specified, but "collected at the POC at four external sites" for method comparison suggests prospective collection in a clinical setting.
• Limit of Blank (LoB): 195 replicate measurements in plasma, 144 replicate measurements in blood.
• Limit of Detection (LoD): 24 replicate measurements on each of six low-level β-hCG blood samples.
• Limit of Quantitation (LoQ): 92 replicate measurements of six low-level β-hCG blood samples.
• Analytical Specificity (Interference): "human plasma" with/without specific interferences; "17 human plasma samples" for HAMA/RF.
• Recovery: "heparinized whole blood and plasma samples from six donors."
• Method Comparison with Predicate Device: 134 samples.
  • Provenance: "Blood samples were collected at the POC at four external sites in heparinized evacuated tubes, and analyzed in duplicate on the i-STAT System. The blood tubes were sent to the laboratory and the plasma portion was separated from the red cells. The plasma portion was tested in duplicate on the i-STAT System and the ARCHITECT system within 6 hours of collection." This indicates a prospective collection for comparison.
• Sample Matrix Comparison: 40 donor samples, each spiked to create 40 levels of hCG.
• Reference Range:
  • 123 apparently healthy, non-pregnant females ≥ 18 and < 40 years.
  • 125 apparently healthy, non-pregnant females > 40 years.
  • Provenance: Specimens from these individuals were "collected and tested at four external sites." This is prospective collection.

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) test, not an imaging device that typically requires expert interpretation for ground truth. The "ground truth" for the performance studies is established by the concentration of β-hCG in the samples, either:

• Known concentrations: Prepared by spiking samples with WHO 5th IS (Recovery studies).
• Reference measurements: From the predicate device (ARCHITECT System) or other reference methods for comparison (Method Comparison, LoB, LoD, LoQ determination using ARCHITECT results for blank plasma/blood), or from CLSI guidelines for establishing reference intervals.
• Clinical context: "Early detection of pregnancy consistent with a β-hCG result above 25 IU/L" (referenced from Tietz NW, Clinical Guide to Laboratory Tests, 4th Ed. 2006).

Therefore, there were no human experts (like radiologists) establishing ground truth in the way conceptualized for AI in medical imaging. The ground truth is biochemical measurement-based.

4. Adjudication method (e.g., 2+1, 3+1, none) for the test set:

Not applicable. As an IVD device measuring a biochemical marker, adjudication by experts for discrepancies in qualitative or quantitative assessment is not a standard practice for establishing ground truth in the way it is for image-based diagnostic systems. The "ground truth" relies on the measured concentration by established reference methods or controlled spiking.

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 test, not an AI-powered image analysis system or a device that directly assists human readers/interpreters. There are no "human readers" in the context of an MRMC study for this type of device.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:

Yes, the studies described are standalone device performance studies. The i-STAT Total β-hCG Test is an automated system that provides quantitative and qualitative results for β-hCG. Its performance characteristics (precision, linearity, limits, specificity, recovery) are evaluated based on its direct output without human intervention for interpretation of the result itself, beyond operating the device and reporting the numerical value. The "algorithm" in this context is the analytical process of the immunoassay and electrochemical detection within the i-STAT system.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

The ground truth used is primarily biochemical measurement and known concentrations:

• Known concentrations: For linearity, recovery, and high dose hook effect studies, samples were spiked to achieve specific, known β-hCG concentrations (often referenced to WHO 5th IS).
• Reference device measurements: For method comparison, the predicate device (Abbott ARCHITECT System Total β-hCG) provided the reference measurements. For LoB/LoD/LoQ determinations, the ARCHITECT system was also used to confirm blank samples.
• Absence of analyte: For LoB, "blank plasma and fresh whole blood that measured < 1.20 IU/L on the ARCHITECT system."
• Clinical criteria/literature: For defining the significance of β-hCG levels (e.g., 25 IU/L for early pregnancy detection, referenced to Tietz NW).
• Expected analyte behavior: For analytical specificity, known interfering substances were spiked in at controlled concentrations.

8. The sample size for the training set:

Not explicitly stated. This device is an in vitro diagnostic test, not a machine learning model that typically relies on a distinct "training set" in the same way as AI software. The development of such devices often involves extensive R&D and optimization using various batches and preparations of reagents and samples, which serves an analogous purpose to training, but it's not generally quantified as a "training set" in regulatory submissions for these types of devices.

9. How the ground truth for the training set was established:

Not applicable in the conventional sense of an AI training set. For the development and optimization of the i-STAT β-hCG test, the ground truth would have been established through:

• Chemical and biological characterization: Precisely prepared calibrators and controls with known concentrations of β-hCG (often traceable to international standards like the WHO 5th IS).
• Cross-validation with established laboratory methods: Using recognized reference methods or existing FDA-cleared predicate devices to confirm the accuracy and performance of the evolving i-STAT assay during its development.
• Internal quality control procedures: Continuous monitoring against known standards and controls during the manufacturing and R&D phases.

Ask a specific question about this device