Search Results

The LYHER® Oral fluid Multi-Drug Test Kit (Cube) is a rapid lateral flow immunoassay for the qualitative detection of d-Amphetamine, d-Methamphetamine, Benzoylecgonine, Morphine, Phencyclidine and Delta-9-Tetrahydrocanabinol in human oral fluid. The test cut-off concentrations and the compounds the tests are calibrated to are as follows:

The single or multi-test panels can consist of the above insted analytes in anycombination, up to a maximum of 6 analytes. The tests provide only a preliminary result. A more specific alternate chemical must be used in order to obtain a confirmed analytical test result. Gas Chromatography/Mass Spectrometry (GC/MS), Liquid Chromatography / Mass Spectrometry (LC/MS) and their tandem mass-spectrometer versions are the preferred confirmatory methods. Careful consideration and judgment should be applied to any drugs of abuse screen test result, particularly when evaluating preliminary positive results.

The LYHER® Oral fluid Multi-Drug Test Kit (Cube) is an immunochromatographic assay that uses a lateral flow system for the qualitative detection of d-Amphetamine, d-Methamphetamine, Benzoylecgonine, Morphine, Phencyclidine and Delta-9-Tetrahydrocannabinol in human oral fluid. The LYHER® Oral fluid Multi-Drug Test Kit (Cube) device consists of a cube device, an oral fluid collection swab and a package insert.

The document describes the analytical performance studies for the LYHER® Oral fluid Multi-Drug Test Kit (Cube), a rapid lateral flow immunoassay. The device is designed to detect d-Amphetamine, d-Methamphetamine, Benzoylecgonine, Morphine, Phencyclidine, and Delta-9-Tetrahydrocannabinol in human oral fluid.

Here's a breakdown of the requested information based on the provided text:

1. A table of acceptance criteria and the reported device performance

The document does not explicitly state "acceptance criteria" as a pass/fail threshold for performance, but rather presents the results of precision-reproducibility studies across various concentrations relative to the cut-off. The performance is summarized by the number of positive and negative results at each concentration.

To construct a table of "acceptance criteria" (inferred as target performance, though not explicitly defined as such) and reported performance, we can focus on the precision-reproducibility studies around the cut-off and the comparison study data. For the comparison study, the implicit acceptance criterion is that the device accurately identifies samples as positive or negative relative to the LC/MS confirmation.

Inferred Acceptance Criteria (Conceptual) and Reported Device Performance

Note on Acceptance Criteria: The document provides raw performance data. For a qualitative immunoassay, the "acceptance criteria" are usually demonstrated by a high rate of negativity well below the cutoff, a high rate of positivity well above the cutoff, and a predictable transition zone around the cutoff concentration. The data presented supports this.

2. Sample size used for the test set and the data provenance

• Test Set Sample Size:

  • Precision-Reproducibility-Cut-Off Study: For each drug, 9 concentration levels were tested. For each concentration, there were 2 runs per day for 30 days, across 3 device lots, and by 3 operators.
    • This means 9 concentrations * 2 runs/day * 30 days * 3 lots * 3 operators = 4860 total tests per drug analyte for the precision study.
    • Each concentration within each operator/lot combination was tested 60 times (2 runs/day * 30 days). So for each specific concentration, lot, and operator, n=60.
  • Comparison Studies (Method Comparison):
    • Negative oral fluid: 360 samples (across all operators and sites for each drug).
    • Positive oral fluid at various ranges:
      • +50% cut off: For D-Amphetamine, 183 samples. For Cocaine, 186 samples. For d-Methamphetamine, 189 samples. For Morphine, 180 samples. For Phencyclidine, 192 samples. For Delta-9-Tetrahydrocannabinol, 195 samples.
    • The total number of samples for the comparison study for each drug is 360 (negatives) + (sum of the positive ranges for that drug). For example, for D-Amphetamine: 360 + 93 + 74 + 180 + 183 = 890 samples. This applies similarly to other drugs.

• Data Provenance: The document states "Method comparison studies for the LYHER Oral fluid Multi-Drug Test Kit(Cube) were performed at three testing sites with three operators at each site." The location of these sites is not explicitly mentioned (e.g., country of origin), but the submitter is based in China. The data origin is prospective as samples were prepared by spiking known concentrations into negative oral fluid for analytical studies.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Ground Truth Establishment: The ground truth for the test set for both the precision and comparison studies was established by Laboratory Confirmed Concentrations, specifically using LC/MS (Liquid Chromatography/Mass Spectrometry). For the precision study, it states: "Each drug concentration was confirmed by LC/MS." For the method comparison studies, it states: "Operators tested the samples using the candidate device and the results were compared to LC/MS results."
• Number of Experts/Qualifications: LC/MS is a laboratory analytical method, not reliant on subjective expert interpretation like radiological imaging. Therefore, there were no "experts" in the sense of human annotators (e.g., radiologists) involved in establishing the ground truth via consensus or adjudication. The "experts" would be the qualified laboratory personnel performing and interpreting the LC/MS results, whose qualifications are implicit given the professional standards for such testing.

4. Adjudication method for the test set

• Since LC/MS is used to establish quantitative drug concentrations, and the device provides a qualitative "positive" or "negative" result based on a defined cutoff, there is no expert adjudication method (like 2+1, 3+1) mentioned or necessary. The device's result is compared directly to the LC/MS result relative to the established cut-off.

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 question is not applicable to this device. This is a qualitative diagnostic test (immunoassay) for drug detection in oral fluid, not an AI-powered image analysis device that assists human readers. The "operators" mentioned are performing the test, not interpreting complex medical images.

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

• This device is a standalone test (a rapid lateral flow immunoassay), essentially an "algorithm only" in the sense that its chemical reactions produce a visual result (line presence/absence). Human interaction involves collecting the sample, applying it to the device, and visually interpreting the presence or absence of test lines. There isn't a separate "algorithm" that operates outside of the device itself to interpret the results. The performance data presented measures the device's inherent analytical accuracy against known concentrations and LC/MS.

7. The type of ground truth used

• The ground truth used for both precision and comparison studies was laboratory-confirmed drug concentrations via LC/MS (Liquid Chromatography/Mass Spectrometry). This is a highly accurate and quantitative analytical method, considered a gold standard for drug detection and quantification.

8. The sample size for the training set

• This device is a lateral flow immunoassay, not a machine learning or AI-based device that requires a "training set" in the computational sense. Its performance is based on the inherent chemical and biological properties of the reagents and test strip design. Therefore, the concept of a "training set" for model development is not applicable. The studies described are analytical validation studies, not AI model training.

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

• As explained in point 8, there is no "training set" for an AI model. The ground truth for the analytical validation (which is analogous to testing the device's inherent design performance) was established by LC/MS confirmation of spiked drug concentrations in oral fluid.

Ask a specific question about this device