FDA 510(k) Search - By Innolitics (SaMD and AI Experts)

AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview

Intended Use

The LZI Cotinine II Enzyme Immunoassay is intended for the quantitative determination of cotinine in human urine at the cutoff value of 200 ng/mL when calibrated against cotinine. The assay is intended as an aid in the detection of cotinine after use or exposure to tobacco products. The assay is designed for prescription use with a number of automated clinical chemistry analyzers.

The semi-quantitative mode is for purposes of (1) enabling laboratories to determine an appropriate dilution of the specimen for confirmation by a confirmatory method such as gas or liquid chromatography/mass spectrometry (GC/MS or LC/MS) or (2) permitting laboratories to establish quality control procedures.

The assay provides only a preliminary analytical result. A more specific alternative chemical method (e.g., gas or liquid chromatography and mass spectrometry) must be used in order to obtain a confirmed analytical result. Clinical consideration and professional judgment should be exercised with any drug of abuse test result, particularly when the preliminary test result is positive.

Device Description

The LZI Cotinine II Enzyme Immunoassay is a homogeneous enzyme immunoassay with readyto-use liguid reagents. The assay is based on competition between drug in the sample and drug labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for a fixed amount of antibody in the reagent. Enzyme activity decreases upon binding to the antibody, and the drug concentration in the sample is measured in terms of enzyme activity. In the absence of drug in the sample, cotinine-labeled G6PDH conjugate is bound to antibody, and the enzyme activity is inhibited. On the other hand, when free drug is present in the sample, antibody would bind to free drug; the unbound cotinine-labeled G6PDH then exhibits its maximal enzyme activity. Active enzyme converts nicotinamide adenine dinucleotide (NAD) to NADH, resulting in an absorbance change that can be measured spectrophotometrically at 340 nm.

The LZI Cotinine II Enzyme Immunoassay is a kit comprised of two reagents, an R1 and R2 which are bottled separately but sold together within the kit.

The R1 solution contains mouse monoclonal anti-cotinine antibody, glucose-6-phosphate (G6P) nicotinamide adenine dinucleotide (NAD), stabilizers, and sodium azide (0.09 %) as a preservative. The R2 solution contains glucose-6-phosphate dehydrogenase (G6PDH) labeled with cotinine in buffer with sodium azide (0.09 %) as a preservative.

AI/ML Overview

The provided text describes the LZI Cotinine II Enzyme Immunoassay, its intended use, and performance characteristics, but it does not describe a study that directly establishes acceptance criteria and directly proves the device meets specific acceptance criteria in a structured manner you requested.

The document is a 510(k) Summary of Safety and Effectiveness, which typically outlines the device's characteristics and demonstrates substantial equivalence to a predicate device. It includes performance data such as precision, linearity, and method comparison (accuracy) with clinical samples, which implicitly serve to show the device is safe and effective for its stated intended use. However, explicit acceptance criteria values that the device must meet are not directly listed in a table format with corresponding proof of meeting them.

I will extract the closest information available to address your request, making inferences where necessary based on the context of common performance expectations for such devices.

Here's a breakdown of the information as requested, largely derived from the "Method Comparison - Clinical Samples" and "Precision" sections.

Acceptance Criteria and Device Performance for LZI Cotinine II Enzyme Immunoassay

Note: The document does not explicitly state "acceptance criteria" in a definitive table format with corresponding proof. Instead, it presents performance study results that are implicitly considered acceptable for demonstrating substantial equivalence. The "Reported Device Performance" below is directly from the study results. The "Acceptance Criteria" are inferred based on the observed performance and typical expectations for diagnostic assays.

1. Table of Acceptance Criteria and Reported Device Performance

Performance Metric	Inferred Acceptance Criterion (Approximate)	Reported Device Performance (from Clinical Samples)	Study Comments
Accuracy (Semi-Quantitative)	High overall agreement with LC/MS, with minimal discrepancies, especially away from the cutoff concentration.	Overall Agreement: ~96.4% for positive, ~98.7% for negative.
Discrepant Samples:

1 sample (128.6 ng/mL LC/MS) was positive by EIA but was between 50% below cutoff and cutoff concentration (100-199.9 ng/mL). The report states it was "discrepant" but the EIA result (positive) aligns with the stated LC/MS value being within the range where a positive result starts to become more likely if it's "near cutoff positive". However, interpreting that specific discrepancy is complex without a clear cut-off definition of "positive" for LC/MS in the table.
1 sample (204.2 ng/mL LC/MS) was negative by EIA but was between cutoff and 50% above cutoff concentration (200-299.9 ng/mL). The report states it was "discrepant". | The accuracy study uses LC/MS as the gold standard. The reported percentages reflect agreement for positive and negative results with respect to the 200 ng/mL cut-off. The discrepancies highlight inherent variability near the cutoff. |
| Accuracy (Qualitative) | High overall agreement with LC/MS, similar to semi-quantitative.| Overall Agreement: ~96.4% for positive, ~98.7% for negative.
Discrepant Samples:
1 sample (128.6 ng/mL LC/MS) was positive by EIA (146.6 mAU) with a qualitative cutoff rate of 126.4 mAU. This sample was LC/MS positive but between 100-199.9 ng/mL, the EIA result was positive.
1 sample (204.2 ng/mL LC/MS) was negative by EIA (93.8 mAU) with a qualitative cutoff rate of 126.4 mAU. This sample was LC/MS positive (204.2 ng/mL) but the EIA result was negative. | Similar to semi-quantitative accuracy, with concordance defined by the qualitative response based on mAU compared to a cutoff rate. |
| Precision (Qualitative) | At 200 ng/mL cutoff, some variability is expected. For concentrations significantly above or below the cutoff, 100% agreement with expected result (positive/negative) is expected. | At 200 ng/mL:
Within Run (N=22): 10 Neg/12 Pos
Total Precision (N=88): 51 Neg/37 Pos
At 250-400 ng/mL: 100% Positive
At 0-150 ng/mL: 100% Negative | This indicates that at the exact cutoff concentration (200 ng/mL), there is expected variability in classification. This is typical for assays and is why results near the cutoff require careful interpretation and often confirmatory testing. |
| Precision (Semi-Quantitative)| Same as qualitative. At 200 ng/mL cutoff, some variability is expected. For concentrations significantly above or below the cutoff, 100% agreement with expected result (positive/negative) is expected. | At 200 ng/mL:
Within Run (N=22): 14 Neg/8 Pos
Total Precision (N=88): 55 Neg/33 Pos
At 250-400 ng/mL: 100% Positive
At 0-150 ng/mL: 100% Negative | The semi-quantitative precision results also show variability at the 200 ng/mL cutoff, with 100% agreement for samples sufficiently above or below the cutoff. This confirms the device's ability to consistently classify samples away from the decision point. |
| Linearity | Percent recovery between 85% and 115% of expected values across the assay's linear range. | Range 100 ng/mL to 1000 ng/mL: Recovery ranging from 97.9% to 110.8%.
At 20 ng/mL: 167.5% recovery.
At 0 ng/mL: N/A (reported 19.3 ng/mL determined). | The device demonstrated good linearity within the critical range for its intended use (100-1000 ng/mL). The higher recovery at 20 ng/mL suggests reduced accuracy at very low concentrations, which is generally not a concern given the 200 ng/mL cutoff. |
| Cross-reactivity | Minimal or no significant cross-reactivity with common substances, particularly structurally unrelated compounds, at physiologically relevant concentrations or concentrations higher than expected. | Cotinine-related compounds: Some cross-reactivity with compounds like (-) Norcotinine (20.00%) and (R, S)-Norcotinine (23.53%).
Structurally Unrelated Pharmacological Compounds: "ND" (Not Detected) at high concentrations (e.g., 100,000 ng/mL) for most tested drugs. Negative results for 0 ng/mL cotinine spiked with these compounds, and correct positive/negative results for cotinine controls. | The device largely demonstrates specificity against a wide panel of common drugs and related substances, indicating minimal interference for its intended purpose. Some expected cross-reactivity with specific cotinine metabolites is noted. |
| Endogenous Interference | No significant interference from common endogenous substances at relevant concentrations. | No significant undesired cross-reactants or endogenous substance interference was observed except for Boric Acid (1000 ng/mL) and Citric Acid (pH 3) (800 ng/mL) which showed interference at ±25% of cutoff and also at ±50% of cutoff. | Most endogenous substances showed no interference. Boric acid and citric acid showed interference, which is important for laboratories to be aware of when testing samples that might contain these substances. |
| Specific Gravity Interference | No interference observed across the tested range. | No interference was observed in samples ranging from 1.005 to 1.028 specific gravity when spiked with cotinine at control concentrations. | This indicates the device is robust to variations in urine specific gravity, which is a common factor in urine drug testing. |
| pH Interference | No major interference observed across the tested pH range. | No major interference between pH 3 to pH 11. | This demonstrates the device's robustness across a wide physiological pH range for urine, an important aspect for clinical utility. |

2. Sample Size Used for the Test Set and Data Provenance

Accuracy (Method Comparison - Clinical Samples):
- Sample Size: 104 unaltered clinical samples.
- Data Provenance: Samples were collected by Lin-Zhi International, Inc. (LZI). The country of origin is not specified, but LZI is based in Santa Clara, CA, USA. The samples are indicated to be "clinical samples," implying they were prospectively collected from human subjects for diagnostic purposes. They are referred to as "unaltered," suggesting they were tested as received without intentional modification. The study is retrospective in the sense that these samples were then tested against a new assay.
Precision:
- Sample Size: For each cotinine concentration, 22 determinations were made within a run, and a total of 88 determinations were made across multiple runs (2 runs/day for 22 days).
- Data Provenance: Samples were prepared by spiking a cotinine standard into a "pool of negative human urine." This indicates a controlled laboratory setting (likely in-house at LZI) rather than clinical samples, and thus not tied to a specific country of origin from patients.
Linearity, Cross-reactivity, Endogenous Interference, Specific Gravity, pH Interference: These studies used spiked samples prepared in a laboratory setting (e.g., "drug free-urine pool," "pool of negative human urine"). The sample sizes for each condition are detailed in the tables (e.g., "10 replicates" for linearity, "replicates" for others).

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

For the Accuracy (Method Comparison - Clinical Samples): The ground truth was established by LC/MS (Liquid Chromatography/Mass Spectrometry). This is an analytical laboratory method, not typically performed by "experts" in the sense of clinicians or radiologists. The result from the LC/MS machine is the ground truth. Therefore, the concept of "number of experts" and their "qualifications" is not applicable here as the ground truth is an objective chemical measurement.

4. Adjudication Method for the Test Set

For the accuracy study, the "ground truth" was established by LC/MS, which is a definitive analytical method. Therefore, no human adjudication method (like 2+1, 3+1 consensus) was used or required for determining the confirmed cotinine concentration. The comparison was directly between the LZI Cotinine II Enzyme Immunoassay results and the LC/MS results.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done. This type of study is typical for imaging devices where human readers interpret images, sometimes with and without AI assistance, to assess AI's impact on human performance. The LZI Cotinine II Enzyme Immunoassay is an in-vitro diagnostic (IVD) device (a laboratory test), not an imaging device that would involve human readers for interpretation in this manner.

6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done

Yes, the performance studies described are essentially standalone evaluations of the LZI Cotinine II Enzyme Immunoassay. The accuracy, precision, linearity, and interference studies assess the analytical performance of the device itself (the "algorithm" in a broad sense, referring to the assay's methodology and reagents) against established reference methods or known concentrations, without requiring human intervention for interpretation beyond operating the automated clinical analyzer and analyzing the data. The results (e.g., ng/mL concentrations or positive/negative classifications) are generated directly by the assay on the automated analyzer.

7. The Type of Ground Truth Used

Primary Ground Truth for Clinical Accuracy: LC/MS (Liquid Chromatography/Mass Spectrometry) for cotinine concentrations. This is a highly specific and sensitive analytical technique, often considered the gold standard for drug confirmation testing.
Ground Truth for Other Studies (Precision, Linearity, Cross-reactivity, etc.): Known concentrations of cotinine or other substances (spiked samples) in negative human urine pools.

8. The Sample Size for the Training Set

The document describes performance studies for device validation. It does not provide information about a "training set" in the context of an AI/machine learning model. This device is an enzyme immunoassay, a chemical assay, not an AI-based diagnostic algorithm that typically undergoes a separate training phase with a large dataset. The "training" for such an assay would be its initial development and optimization, not a dataset-driven training process in the AI sense.

9. How the Ground Truth for the Training Set Was Established

As explained in point 8, the concept of a "training set" for an AI/machine learning model is not applicable to this enzyme immunoassay device. Therefore, how its ground truth was established is not relevant here. The development of the assay's chemical reagents and methodology would have involved internal validation and optimization, but not a "training set" in the context of AI.

§ 862.3220 Carbon monoxide test system.