Search Filters

Search Results

Found 4 results

510(k) Data Aggregation

    Why did this record match?
    Applicant Name (Manufacturer) :

    APPLIED DNA TECHNOLOGIES, INC.

    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Applied DNA Technologies Bionexia™ DOA Panels are rapid chromatographic immunoassays for the qualitative and simultaneous detection of one to thirteen of the following drugs in a variety of combinations in human urine. The designed cutoff concentrations and direct calibrator for these drugs are as follows:

    AnalyteAbbreviationCalibratorCutoff Concentration
    AmphetamineAMPAmphetamine1000 ng/ml
    BarbiturateBARSecobarbital300 ng/ml
    BenzodiazepinesBZOOxazepam300 ng/ml
    CocaineCOCBenzoylecgonine300 ng/ml
    MarijuanaTHC11-nor- $\Delta^{9}$ -THC9-COOH50 ng/ml
    MethamphetamineMETMethamphetamine1000 ng/ml
    MethadoneMTDMethadone300 ng/ml
    MorphineMORMorphine2000 ng/ml
    MorphineMORMorphine300 ng/ml
    PhencyclidinePCPPhencyclidine25 ng/ml
    NortriptylineNORNortriptyline1000 ng/ml
    EcstasyMDMA3,4-Methylenediioxy-MET500 ng/ml
    BuprenorphineBUPBUP-3-D-Glucuronide10 ng/ml
    EDDPEDDP2-ethylidene-1,5-dimethyl-
    3,3-diphenylpyrrolidine100 ng/ml

    For health care professionals use including professionals at point of care sites (POC) to assist in the determination of drug compliance.

    This assay provided only a preliminary analytical test result. A more specific alternative chemical method must be used in order to obtain a confirmed analytical result. Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS) are the preferred confirmatory method.

    Clinical consideration and professional judgment should be applied to any drug of abuse test result, particularly when preliminary positive results are indicated.

    Device Description

    One-step, colloidal gold based chromatographic immunoassay for the rapid, qualitative detection of Marijuana, Cocaine, Phencyclidine, Morphine, Methamphetamine, Methadone, Amphetamine, Barbiturates, Benzodiazepines, Nortriptyline, Ecstasy, Buprenorphine and Methadone metabolite -EDDP, in human urine.

    AI/ML Overview

    This document describes the validation of the Bionexia™ Single and Multi-Strip Cassette/Dipstick DOA Screen Panels for detecting various drugs of abuse in human urine.

    1. Acceptance Criteria and Reported Device Performance

    The acceptance criteria are implied by the reported "Overall Agreement" percentages, which demonstrate the device's accuracy compared to confirmatory methods (GC/MS or LC/MS). While explicit numerical acceptance criteria for each analyte are not stated, the consistently high agreement rates (all above 93%) suggest a benchmark for acceptable performance.

    AnalyteReported Agreement (Previously Approved Panels)Reported Agreement (Additional Panels)
    Amphetamine (AMP)98.1%N/A
    Barbiturate (BAR)98.0%N/A
    Benzodiazepines (BZO)93.9%N/A
    Cocaine (COC)98.2%N/A
    Methamphetamine (MET)98.3%N/A
    Morphine (MOR)98.1%97.9% (for 300 ng/ml cutoff)
    Methadone (MTD)98.1%N/A
    Phencyclidine (PCP)98.9%N/A
    Nortriptyline (NOR)96.8%N/A
    Marijuana (THC)97.5%N/A
    MDMAN/A100.0%
    EDDPN/A100.0%
    Buprenorphine (BUP)N/A100.0%

    Note: For analytes listed in both tables, the 'Previously Approved Panels' table likely refers to clinical specimen correlation against GC/MS/LC/MS, while the 'Additional Panels' table further breaks down agreement by concentration range for newer analytes. The morphine (MOR) entry appears twice with different cutoffs, 2000 ng/ml in the first table and 300 ng/ml in the second.

    2. Sample Sizes and Data Provenance

    Previously Approved Panels (Table 1):

    The sample sizes for the test set vary by analyte and are provided as the denominator in the agreement percentages. For example:

    • AMP: 103 samples
    • BAR: 98 samples
    • BZO: 99 samples
    • COC: 110 samples
    • MET: 115 samples
    • MOR (2000 ng/ml): 105 samples
    • MTD: 105 samples
    • PCP: 94 samples
    • NOR: 95 samples
    • THC: 122 samples

    Additional Panels (Table 2):

    For each drug, the total number of samples is the sum of all categories (No Drug present, Negative, Near-Cutoff Negative, Near-Cutoff Positive, High Positive). For example:

    • MDMA: 35 + 7 + 9 + 12 + 63 = 126 samples
    • EDDP: 38 + 0 + 4 + 1 + 65 = 108 samples
    • BUP: 35 + 1 + 5 + 0 + 72 = 113 samples
    • MOR (300 ng/ml): 35 + 0 + 12 + 2 + 50 = 99 samples

    Data Provenance: The data is derived from "blind-labeled clinical specimen correlation study" and "blind-labeled spiked control studies including point-of-care site study." This indicates the data is retrospective (clinical specimens) and prospective (spiked control studies). The country of origin for the data is not specified, but the submission is to the U.S. FDA, suggesting the studies were conducted to U.S. regulatory standards.

    3. Number of Experts and Qualifications for Ground Truth

    The document does not mention the use of human experts to establish ground truth for the test set. Instead, the ground truth is established by "Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS) methodology," which are analytical chemical methods.

    4. Adjudication Method for the Test Set

    Since the ground truth is established by GC/MS or LC/MS, there is no human adjudication method (e.g., 2+1, 3+1) described or implied for the test set. The device results are directly compared to the analytical results from the confirmatory methods.

    5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

    There is no MRMC comparative effectiveness study mentioned in the document. The study focuses on the performance of the Bionexia™ device compared to predicate devices and confirmatory chemical methods. There is no information provided regarding human readers' performance with or without AI assistance.

    6. Standalone Performance Study

    Yes, a standalone performance study was conducted. The tables provided (Table 1 and Table 2) detail the Bionexia™ device's performance (algorithm only, as it's an immunoassay) against the gold standard methods (GC/MS or LC/MS) or predicate devices. This represents the device's performance without a human-in-the-loop directly influencing the test result interpretation (though a healthcare professional would interpret the device's output).

    7. Type of Ground Truth Used

    The type of ground truth used is confirmatory analytical chemical methods, specifically Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS). The document explicitly states these are "the preferred confirmatory method."

    8. Sample Size for the Training Set

    The document does not provide information regarding a separate "training set" or its sample size. The studies described are performance evaluations, implying these samples were used for validation rather than algorithm training. For a traditional immunoassay, there wouldn't typically be an "algorithm training set" in the same sense as machine learning. The device's characteristics are determined during its development and manufacturing process, then validated against known samples.

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

    As there is no mention of a training set in the context of algorithm development, there is no information on how its ground truth was established. For the validation/test sets, as described in point 7, the ground truth was established using GC/MS or LC/MS.

    Ask a Question

    Ask a specific question about this device

    K Number
    K070921
    Date Cleared
    2007-06-29

    (88 days)

    Product Code
    Regulation Number
    862.1155
    Reference & Predicate Devices
    N/A
    Why did this record match?
    Applicant Name (Manufacturer) :

    APPLIED DNA TECHNOLOGIES, INC.

    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Applied DNA Technologies Bionexia™ hCG Pregnancy Cassette and Dipstick Tests are rapid chromatographic immunoassays for the visual, qualitative detection of human chorionic gonadotropin (hCG) in urine to help in the early determination of pregnancy.

    The test kits are for health care professionals use including professionals at physician's office labs (POLs).

    For a final Diagnosis of pregnancy, a more specific alternative clinical method must be used to obtain a confirmed analytical result.

    Device Description

    The Bionexia™ hCG Pregnancy Test will be distributed in both Cassette and Dipstick formats. Each test reagent strip contains mouse monoclonal anti-a-hCG antibody coated membrane and a dried chemical pad containing mouse monoclonal anti-ß-hCG anybody colloidal gold conjugate. The control antibodies are goat anti-mouse IgG.

    AI/ML Overview

    Here's a breakdown of the acceptance criteria and study details for the Bionexia™ hCG Pregnancy Cassette and Dipstick Tests, based on the provided 510(k) summary:

    1. Table of Acceptance Criteria and Reported Device Performance

    Acceptance Criteria / Performance MetricTarget / CriterionReported Device Performance
    Accuracy / Comparison StudySubstantial equivalence to predicate device (ACON One Step Pregnancy Urine Test)100% agreement with the predicate device (ACON One Step hCG Test Card) in 95 clinical urine specimens. (48 positive, 47 negative). Re-evaluation of 20 negative and 29 positive specimens at a third site also showed similar results.
    SensitivityDetection of ß-Human Chorionic Gonadotropin at a concentration of 20 mIU/ml or greater in urine specimen. Standardized to WHO Fourth International Standard 75/589.The Bionexia™ hCG Pregnancy Test detects ß-Human Chorionic Gonadotropin at a concentration of 20 mIU/ml or greater.

    The provided table shows (at 20 mIU/ml): 100% positive results across 4 lots (100 out of 100 specimens detected as positive at 20 mIU/ml).

    For levels below 20 mIU/ml, there's a graded response, suggesting the 20 mIU/ml threshold is met (e.g., at 17.5 mIU/ml, 92% positive; at 15 mIU/ml, 56% positive). |
    | Specificity (Cross-reactivity) | No cross-reaction with interfering substances at specified levels. | hCG (20 mIU/ml): 100% Non-cross-reactivity (This seems to imply 100% specific for hCG at 20mIU/ml, or perhaps the table header is misleading and it refers to 100% detection of hCG, but the context of "cross-reactivity" suggests it means no cross-reactivity with other substances at the hCG levels tested for specificity. Given the other rows, it's more likely this indicates it correctly identifies hCG and isn't cross-reacting with other substances as if it were hCG.).
    hLH (300 mIU/ml): >500% (likely 0% cross-reactivity, implies at 300mIU/ml hLH, the device gives a negative result).
    hFSH (1000 mIU/ml): >5.000% (likely 0% cross-reactivity).
    hTSH (1.000 ulUmL): >5% (likely 0% cross-reactivity).
    (Interpretation of "% Non-cross-reactivity" and the values 500% and 5.000% is ambiguous; it most likely means no detectable cross-reactivity at these high concentrations, where a negative result is desired.) |
    | Reproducibility | Demonstrated consistency across multiple tests/lots. | Results demonstrated substantial equivalency with the predicate device. (Specific quantitative results not provided in the summary, but implied through this statement). |
    | Safety and Effectiveness | Demonstrated as safe and effective. | Demonstrated as safe and effective in detecting hCG in urine. |

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

    • Accuracy/Comparison Study (Test Set):
      • Sample Size: 95 clinical urine specimens (48 positive, 47 negative). An additional 49 specimens (20 negative, 29 positive) were re-evaluated at a third site.
      • Data Provenance: Clinical specimen correlation study, including POLs (Physician's Office Labs) site study. This indicates the data is prospective from clinical settings. The country of origin is not explicitly stated in the provided text, but the submitter's address is in San Diego, CA, USA, implying the study was likely conducted within the USA.
    • Sensitivity/Cross-reactivity Study (Test Set):
      • Sample Size: For sensitivity, 100 specimens per level per lot, across 4 lots (totaling 400 specimens tested across the concentration curve). For cross-reactivity, the specific number of spiked samples is not given, but results are provided for various substances at specified levels.
      • Data Provenance: This appears to be a laboratory-based study using spiked controls, rather than clinical specimens for this specific component.

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

    • The document does not explicitly state the number of experts or their specific qualifications for establishing ground truth.
    • For the accuracy/comparison study, the comparison is made against a "predicate device" (ACON One Step Pregnancy Urine Test). The "ground truth" for these clinical samples is implicitly the result obtained from the predicate device, as the study aims to show agreement with it.
    • For the sensitivity and cross-reactivity studies, the ground truth is established by the known concentration of spiked hCG or other substances.

    4. Adjudication Method for the Test Set

    • The document does not describe an adjudication method for the test set. Given that the comparison study is against a predicate device, it's likely that the predicate device's result was considered the reference, and discrepancies would have been examined, but the process is not detailed.

    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

    • No, an MRMC comparative effectiveness study was not done. This device is a rapid diagnostic test (lateral flow immunoassay) for visual reading, not an AI-powered diagnostic imaging device that assists human readers.

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

    • Yes, a standalone performance was done. This device is intended to be read visually by healthcare professionals, meaning its performance is inherently a "standalone" interpretation of the test result without an AI algorithm. The performance evaluation directly assesses the device's ability to produce accurate results.

    7. The Type of Ground Truth Used

    • Accuracy/Comparison Study: The ground truth for clinical specimens was derived from the results of the predicate device (ACON One Step Pregnancy Urine Test).
    • Sensitivity and Specificity (Cross-reactivity) Studies: The ground truth was based on known concentrations of spiked analytes (hCG) and known concentrations of potential interfering substances. This is a form of laboratory-controlled ground truth.

    8. The Sample Size for the Training Set

    • The provided document does not mention a "training set" in the context of machine learning or AI. This is a point-of-care diagnostic device, not an AI/ML algorithm. Therefore, there is no training set in that sense. The studies described are performance validation studies.

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

    • As there is no training set for an AI/ML algorithm, this question is not applicable to the Bionexia™ hCG Pregnancy Test.
    Ask a Question

    Ask a specific question about this device

    K Number
    K071030
    Date Cleared
    2007-06-29

    (79 days)

    Product Code
    Regulation Number
    862.1155
    Reference & Predicate Devices
    N/A
    Why did this record match?
    Applicant Name (Manufacturer) :

    APPLIED DNA TECHNOLOGIES, INC.

    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Applied DNA Technologies Bionexia™ hCG Pregnancy Serum/Urine Cassette and Dipstick Tests are rapid chromatographic immunoassays for the visual, qualitative detection of human chorionic gonadotropin (hCG) in serum or urine specimen to help in the early determination of pregnancy. The test kits are for health care professionals use including professionals at physician's office labs (POLs). For a final Diagnosis of pregnancy, a more specific alternative clinical method must be used to obtain a confirmed analytical result.

    Device Description

    The Bionexia™ hCG Pregnancy Serum/Urine Test are distributed in both Cassette and Dipstick formats. Each test reagent strip contains mouse monoclonal anti-u-hCG antibody coated membrane and a dried chemical pad containing mouse monoclonal anti-ß-hCG anybody colloidal gold conjugate. The control antibodies are goat anti-mouse IgG.

    AI/ML Overview

    Here's a breakdown of the acceptance criteria and the study details for the Bionexia™ hCG Pregnancy Serum/Urine Cassette and Dipstick Tests, based on the provided 510(k) summary:

    1. Table of Acceptance Criteria and Reported Device Performance

    Acceptance CriterionReported Device Performance (Bionexia™ hCG Tests)
    Accuracy (Serum)100% agreement with predicate device (ACON One Step hCG Urine/Serum Test Card) for 95 clinical serum specimens.
    Accuracy (Urine)100% agreement with predicate device (ACON One Step hCG Urine/Serum Test Card) for 94 clinical urine specimens.
    SensitivityDetects serum or urinary hCG at a concentration of 20 mIU/ml or greater.
    Cross-reactivityNo cross-reaction observed for hLH (300 mIU/ml), hFSH (1000 mIU/ml), and hTSH (1,000 µIU/mL) at 0 mIU/ml hCG. 100% non-cross-reactivity for hCG at 20 mIU/ml (this is essentially a positive control).

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

    • Accuracy/Comparison Study (Test Set):

      • Serum: 95 clinical serum specimens.
      • Urine: 94 clinical urine specimens.
      • Additional evaluation: 20 negative and 29 positive serum & urine specimens (total 49) were re-evaluated at a third site.
      • Data Provenance: Clinical specimens. The study was conducted at two Physician's Office Labs (POLs) sites, with additional re-evaluation at a third site. The country of origin is not explicitly stated, but the submission is to the FDA in the USA. The data appears to be prospective in the sense that the new device was tested on these clinical specimens.
    • Sensitivity and Cross-reactivity Study (Test Set):

      • Sensitivity: Standardized to the WHO Fourth International Standard 75/589, implying a controlled, possibly spiked, sample set. Specific sample count not provided, but it states detection at 20mIU/ml or greater.
      • Cross-reactivity: Evaluated at 0 mIU/ml (negative) and 20 mIU/ml (positive) hCG specimens. Specific sample count for each cross-reactant not provided, but it implies tests performed with each substance at the specified level. Likely spiked control studies.
    • Reproducibility Study (Test Set):

      • Serum Controls: 100 tests (25 tests across 4 lots) for each hCG concentration level (0, 10, 12.5, 15, 17.5, 20, 25, 30, 35, 40, 100 mIU/ml).
      • Urine Controls: 100 tests (25 tests across 4 lots) for each hCG concentration level (0, 10, 12.5, 15, 17.5, 20, 25, 30, 35, 40, 100 mIU/ml).
      • Data Provenance: This appears to be spiked control studies using prepared control samples rather than clinical specimens.

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

    • The document implies that the predicate device (ACON One Step Pregnancy Urine/Serum Test) results were used as the reference point (ground truth) for the accuracy/comparison study. It doesn't explicitly mention external experts adjudicating the Bionexia™ results against a separate "true" ground truth. The comparison is against an already legally marketed and accepted device.
    • For the POLs site study, it is implied that healthcare professionals (the intended users) performed the tests and interpreted the results. The specific number or qualifications of these professionals are not detailed, but they are referred to as "health care professionals use including professionals at physician's office labs (POLs)."

    4. Adjudication Method for the Test Set

    • The primary method for the accuracy/comparison study was a direct comparison to the predicate device. A "blind-labeled" approach was used (meaning the testers likely didn't know which sample was which device), but there's no mention of a formal adjudication method (like 2+1 or 3+1 consensus) for discrepancies if they arose between the new device and the predicate. Given the 100% agreement, no adjudication would have been necessary in those specific cases.

    5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size of How Much Human Readers Improve With AI vs Without AI Assistance

    • No, an MRMC comparative effectiveness study was not done. This document describes the performance of a rapid in-vitro diagnostic (IVD) test, not an AI-assisted diagnostic tool. Therefore, the concept of human readers improving with AI assistance is not applicable here. The device itself performs the detection, with visual interpretation by the healthcare professional.

    6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

    • Yes, a standalone performance was done for the device. The accuracy, sensitivity, cross-reactivity, and reproducibility studies demonstrate the performance of the Bionexia™ hCG Pregnancy Tests themselves, irrespective of the human factor beyond the visual interpretation of the test line. The device is designed for visual qualitative detection, meaning the "algorithm" is inherent in the chemical reactions and visual indicators. The results presented are the device's performance in detecting hCG.

    7. The Type of Ground Truth Used

    • Accuracy/Comparison Study: The ground truth was established by the results of a legally marketed predicate device (ACON One Step Pregnancy Urine/Serum Test). This serves as a comparative ground truth.
    • Sensitivity: The ground truth was based on the WHO Fourth International Standard 75/589 for hCG concentration, which is a standardized reference.
    • Cross-reactivity and Reproducibility: Ground truth was established by using controlled, spiked samples with known concentrations of hCG and other related hormones/substances.

    8. The Sample Size for the Training Set

    • The document does not explicitly describe a "training set" in the context of machine learning or AI. For IVD devices like this, the development process typically involves internal testing and optimization (which could be considered analogous to training data for the device's design), but this is not typically reported as a formal "training set" with specific numbers in regulatory summaries. The data provided focuses on validation (test set) performance.

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

    • As a traditional IVD device, there isn't a "training set ground truth" in the AI sense. The development of the device (choosing antibodies, membrane materials, concentrations, etc.) would be guided by established biochemical principles and internal testing using known hCG standards and samples with confirmed values, which serves a similar function to establishing "ground truth" during product development. However, these specific details of the development process and internal testing are not provided in this 510(k) summary, which focuses on validation data for regulatory submission.
    Ask a Question

    Ask a specific question about this device

    K Number
    K061005
    Date Cleared
    2006-10-04

    (176 days)

    Product Code
    Regulation Number
    862.3100
    Reference & Predicate Devices
    Why did this record match?
    Applicant Name (Manufacturer) :

    APPLIED DNA TECHNOLOGIES, INC.

    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    Applied DNA Technologies ACCUSTEP DOA Panels are rapid chromatographic The immunoassays for the qualitative and simultaneous detection of one to ten of the following drugs in a variety of combinations in human urine. The designed cutoff concentrations and direct calibrator for these drugs are as follows:

    AnalyteAbbreviationCalibratorCutoff Concentration
    AmphetamineAMPAmphetamine1000 ng/ml
    BarbiturateBARSecobarbital300 ng/ml
    BenzodiazepinesBZOOxazepam300 ng/ml
    CocaineCOCBenzoylecgonine300 ng/ml
    MarijuanaTHC11-nor- $\Delta^9$ -THC9-COOH50 ng/ml
    MethamphetamineMETMethamphetamine1000 ng/ml
    MethadoneMTDMethadone300 ng/ml
    MorphineMORMorphine2000 ng/ml
    PhencyclidinePCPPhencyclidine25 ng/ml
    NortriptylineNORNortriptyline1000 ng/ml

    These test kits are intended for health care professional use only.

    This assay provided only a preliminary analytical test result. A more specific alternative chemical method must be used in order to obtain a confirmed analytical result. Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS) are the preferred confirmatory method.

    Clinical consideration and professional judgment should be applied to any drug of abuse test result, particularly when preliminary positive results are indicated.

    Device Description

    One-step, colloidal gold based chromatographic immunoassay for the rapid, qualitative detection of Marijuana. Cocaine, Phencyclidine, Morphine, Methamphetamine, Methadone, Amphetamine, Barbiturates, Benzodiazepines and Nortriptyline, a Tricyclic Antidepressant, in human urine.

    AI/ML Overview

    The ACCUSTEP Single and Multi-Strip Cassette/Dipstick DOA Screen Panels are rapid chromatographic immunoassays designed for the qualitative and simultaneous detection of various drugs in human urine. The study presented supports the device's substantial equivalence to predicate devices and GC/MS methodology.

    Here's an analysis of the acceptance criteria and study details:

    1. Table of Acceptance Criteria and Reported Device Performance

    The acceptance criteria are implied by the "Overall Agreement" reported against GC/MS analysis. While specific numerical acceptance criteria (e.g., "must achieve >95% overall agreement") are not explicitly stated, the presented performance data demonstrates high agreement with the gold standard.

    AnalyteAcceptance Criteria (Implied)Reported Device Performance (Overall Agreement with GC/MS)
    AmphetamineHigh agreement with GC/MS98.1% (101/103)
    BarbiturateHigh agreement with GC/MS98.0% (96/98)
    BenzodiazepinesHigh agreement with GC/MS93.9% (93/99)
    CocaineHigh agreement with GC/MS98.2% (108/110)
    MethamphetamineHigh agreement with GC/MS98.3% (113/115)
    MorphineHigh agreement with GC/MS98.1% (103/105)
    MethadoneHigh agreement with GC/MS98.1% (103/105)
    PhencyclidineHigh agreement with GC/MS98.9% (93/94)
    NortriptylineHigh agreement with GC/MS96.8% (92/95)
    Marijuana (THC)High agreement with GC/MS (or >92% correlation with predicate)97.5% (119/122)

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

    The sample sizes for the test set vary by analyte, ranging from 94 to 122 clinical urine specimens.

    • AMP: 103 samples (48 positive, 55 negative)
    • BAR: 98 samples (46 positive, 52 negative)
    • BZO: 99 samples (43 positive, 56 negative)
    • COC: 110 samples (56 positive, 54 negative)
    • MET: 115 samples (63 positive, 52 negative)
    • MOR: 105 samples (41 positive, 64 negative)
    • MTD: 105 samples (51 positive, 54 negative)
    • PCP: 94 samples (46 positive, 48 negative)
    • NOR: 95 samples (38 positive, 57 negative)
    • THC: 122 samples (62 positive, 60 negative)

    The data provenance is stated as "blind-labeled clinical specimen correlation study" using "clinical urine specimens." The country of origin is not specified, but the submission is to the US FDA, implying that the specimens would be relevant to the US population or a general population. This is a retrospective study using previously collected clinical samples.

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

    The ground truth for the test set was established using Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS) as the "preferred confirmatory method." These are analytical chemical methods, not human expert evaluations for establishing ground truth in this context. Therefore, human experts were not used to establish the ground truth for the test set; rather, a laboratory analytical method was used.

    4. Adjudication Method for the Test Set

    No adjudication method for the test set is mentioned, as the ground truth was established by analytical chemical methods (GC/MS or LC/MS), not by human expert consensus that would require adjudication.

    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 device is a rapid diagnostic immunoassay for drug detection, not an AI-powered image analysis system or a device that directly assists human readers in interpreting complex cases. Therefore, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance is not applicable and was not performed.

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

    Yes, the performance study directly evaluates the "ACCUSTEP DOA Screening Panels vs. GC/MS Analysis," which represents the standalone performance of the device without human interpretation or intervention in the analytical result. The device itself produces a "qualitative and simultaneous detection" result, which is then compared to the GC/MS result. Health care professionals interpret this result, but the study focuses on the device's accuracy in detecting the drug.

    7. The Type of Ground Truth Used

    The type of ground truth used is analytical chemical confirmation via Gas Chromatography / Mass Spectrometry (GC/MS) or Liquid Chromatography / Mass Spectrometry (LC/MS).

    8. The Sample Size for the Training Set

    The document does not explicitly mention a "training set" or its sample size. This is typical for immunoassay submissions where the device's analytical performance is evaluated rather than an AI model's learning phase. The "performance characteristics... were evaluated in the laboratory settings and in the blind-labeled clinical specimen correlation study" implies an evaluation of the final product, not a training phase.

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

    As no training set is explicitly mentioned for an AI model, the method for establishing its ground truth is not applicable. The device's underlying immunoassay technology does not involve machine learning that requires a separate training set in the conventional sense.

    Ask a Question

    Ask a specific question about this device

    Page 1 of 1