Search Results
Found 6 results
510(k) Data Aggregation
(105 days)
The ARK™ Methotrexate Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of Methotrexate in human serum or automated clinical chemistry analyzers. The measurements obtained are used in monitoring levels of methotrexate to help ensure appropriate therapy.
The ARK™ Methotrexate Calibrator is intended for use in calibration of the ARK Methotrexate Assay.
The ARKTM Methotrexate Control is intended for use in quality control of the ARK Methotrexate Assay.
Specimens from patients who have received glucarpidase (carboxypeptidase G2) as a high dose methotrexate rescue therapy should not be tested with the ARK Methotrexate Assay.
The ARK Methotrexate Assay is a homogeneous immunoassay based on competition between drug in the specimen and Methotrexate labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay.
The ARK Methotrexate Assay consists of reagents R1 anti-Methotrexate polyclonal antibody with substrate and R2 Methotrexate labeled with bacterial G6PDH enzyme. The ARK Methotrexate Calibrator consists of a six-level set to calibrate the assay, and the ARK Methotrexate Control consists of a six-level set used for quality control of the assay (tri-level calibration range set and tri-level high range set). The ARK Methotrexate Dilution Buffer is equivalent to zero calibrator (Calibrator A).
Here's a breakdown of the acceptance criteria and the studies that prove the device meets them, based on the provided text:
Acceptance Criteria and Reported Device Performance
| Acceptance Criteria Category | Specific Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LoQ) | LoB < Result < LoQ: report "analyte detected; concentration < LoQ" Result ≥ LoQ: report the result as measured | LoB: 0.01 µmol/L LoD: 0.02 µmol/L LoQ: 0.04 µmol/L (LoQ-2 SD > LoD met) |
| Accuracy (Analytical Recovery) | Not explicitly stated as a numerical criterion in the "Acceptance Criteria" section for each data point, but typically implied to be close to 100% recovery. | Mean percentage recovery: 102.1% (Individual recoveries ranged from 98.3% to 111.1%) |
| Linearity | Percent difference was ±10% between the predicted 1st and 2nd order regressed values for concentrations >0.10 µmol/L or ±0.01 µmol/L at concentrations ≤ 0.10 µmol/L. | All observed differences for concentrations >0.10 µmol/L were within ±10% (-2.3% to 4.8%).All observed differences for concentrations ≤ 0.10 µmol/L were within ±0.01 µmol/L (-0.010 µmol/L to -0.004 µmol/L). The highest theoretical concentration tested was 1.30 µmol/L. Regression of assayed methotrexate concentrations was linear throughout the range of 2 to 1200 µmol/L when proportionally diluted. |
| Assay Range | Not explicitly stated as an acceptance criterion, but the functional range. | Measurement range: 0.04 - 1.20 µmol/L |
| Method Comparison | Not explicitly stated as a numerical acceptance criterion (e.g., minimum correlation coefficient or confidence intervals for slope/intercept), but comparison to a predicate device is expected. | Range 0.04 to 1.19 µM (102 samples): Slope: 1.00 (95% CI: 1.00 to 1.02) y-intercept: 0.01 (95% CI: 0.00 to 0.01) Correlation Coefficient (r²): 0.978 (95% CI: 0.968 to 0.985)Range 0.04 to 1440 µM (147 samples, including diluted): Slope: 0.99 (95% CI: 0.96 to 1.00) y-intercept: 0.01 (95% CI: 0.01 to 0.01) Correlation Coefficient (r²): 0.998 (95% CI: 0.997 to 0.998) |
| Precision | <10% total CV at >0.10 µmol/L SD ≤0.01 at ≤0.10 µmol/L | All control and patient pool samples met the criteria: - ARK Methotrexate Control (MID, HIGH, 5, 50, 500 µmol/L): Total CVs ranged from 3.8% to 7.0%. - ARK Methotrexate Control (LOW, 0.06 µmol/L): Total SD was 0.007, which is ≤0.01. - Patient Pool (MID, HIGH, 5, 50, 500 µmol/L): Total CVs ranged from 5.3% to 7.2%. - Patient Pool (LOW, 0.07 µmol/L): Total SD was 0.008, which is ≤0.01. |
| Interfering Substances | Not substantially affected by tested endogenous substances. | Measurement of methotrexate was not substantially affected by clinically high concentrations of Albumin, Bilirubin (conjugated & unconjugated), Cholesterol, Gamma-Globulin, Hemoglobin, Intralipid®, Rheumatoid Factor, Triglycerides, and Uric Acid. |
| Specificity - Crossreactivity | < 0.07% crossreactivity with 7-Hydroxymethotrexate. | 7-Hydroxymethotrexate: ≤ 0.07% crossreactivity. DAMPA: 64.3% to 100% crossreactivity (significant). Triamterene: Crossreactivity ranged from 1.85% to 3.32%. Trimethoprim: Crossreactivity ranged from 0.12% to 0.54%. Folate analogs and other compounds (e.g., Adriamycin, Cyclophosphamide): ≤ 0.01% crossreactivity at ≥ 1000 µmol/L. |
| Anticoagulants | No significant difference in recovery between serum and plasma. | No significant difference found in methotrexate recovery between serum and plasma samples. |
| Sample Stability | Acceptable stability under various storage conditions. | Stable frozen (at least 15 months), 48 hours at room temperature, refrigerated (at least 21 days), and after three (3) successive freeze/thaw cycles. |
| On-Board Stability | Acceptable stability of calibration curve and reagents on the analyzer. | Calibration curve effective up to at least 19 days. Reagents effective for up to at least 25 days in analyzer-specific containers. In-use stability of calibrator and controls also demonstrated. |
Study Details
2. Sample sizes used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective):
- Limit of Quantitation (LoQ):
- LoB, LoD: N = 60
- LoQ: N = 40
- Data provenance: Not specified, but generally analytical studies like this are conducted with prepared samples in a laboratory setting (prospective).
- Accuracy: Concentrated methotrexate drug added into human serum negative for methotrexate. Six replicates of each sample were assayed.
- Sample size: 5 target concentrations, each with 6 replicates (30 assays).
- Data provenance: Not specified, but likely prospective, prepared samples in a lab.
- Linearity: Dilutions of a 1.30 µmol/L serum sample and proportional dilutions of samples between 2 and 1200 µmol/L in pooled human serum.
- Sample size: 13 different concentrations were tested for the initial linearity study. The second linearity study for higher concentrations (2-1200 µmol/L) stated "samples containing methotrexate... were prepared proportionally," implying multiple samples across this range.
- Data provenance: Not specified, but likely prospective, prepared samples in a lab.
- Method Comparison:
- Sample size: 102 specimens within the measurement range (0.04 to 1.19 µM) and 147 specimens including those above the measurement range requiring dilution (total range 0.04 to 1440 µM).
- Data provenance: Not specified, but these are likely clinical samples (human serum/plasma), which could be retrospective or prospective. The document does not specify country of origin.
- Precision:
- Sample size: Each of the 6 control levels and 6 patient pool levels were assayed in quadruplicate, twice a day for 20 days. This means 4 * 2 * 20 = 160 measurements per level.
- Data provenance: Not specified, but likely prospective, prepared controls and pooled patient samples in a lab.
- Interfering Substances: Multiple substances tested, each with samples at ~0.05 and ~0.50 µmol/L methotrexate.
- Sample size: Each interfering substance was tested at two methotrexate levels, likely with duplicates or triplicates, plus control samples. The specific number of individual assays is not given but implies a significant number of tests.
- Data provenance: Not specified, but likely prospective, prepared samples in a lab setting.
- Specificity (Crossreactivity):
- Sample size: Two concentrations of methotrexate (0.05 and 0.50 µmol/L) were tested with triamterene and trimethoprim. Various other compounds and folate analogs were tested at high concentrations (e.g., ≥ 1000 µmol/L), presumably just once for cross-reactivity.
- Data provenance: Not specified, but likely prospective, prepared samples in a lab setting.
- Anticoagulants:
- Sample size: "Studies were conducted," implying multiple samples, but no specific number is given.
- Data provenance: Not specified, but likely prospective, prepared samples in a lab setting.
- Sample Stability:
- Sample size: Not specified, but involved human specimens.
- Data provenance: Not specified, but likely prospective.
- On-Board Stability:
- Sample size: Not specified, but utilized a calibration curve and reagents over time.
- Data provenance: Not specified, but likely prospective, lab-based testing.
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 assay for quantitative determination of methotrexate. The "ground truth" for such assays is typically established by:
- Reference materials: Certified reference standards for methotrexate concentration (used in Accuracy, LoQ, Linearity, Crossreactivity).
- Predicate device comparison: The "gold standard" or accepted method (Fluorescence Polarization Immunoassay - FPIA in this case, specifically the Abbott TDx®/TDxFLx® Methotrexate II assay) for method comparison.
Therefore, the concept of "experts" establishing ground truth as it applies to image interpretation (like radiologists) or pathological diagnosis is not directly relevant here. The ground truth is analytical and based on established methodology and certified materials.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set:
Not applicable. Adjudication methods like 2+1 or 3+1 are typically used in clinical studies where subjective human interpretation (e.g., reading medical images) is involved and discrepancies between readers need to be resolved to establish a consensus ground truth. For in vitro diagnostic device performance studies, the results are quantitative and compared against analytical or predicate methods, not subjective expert consensus.
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 device is an in vitro diagnostic assay, not an AI-powered image analysis tool or a device requiring human "readers" in the context of an MRMC study. Its performance is evaluated analytically, not by assessing human performance improvement with or without AI.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
Yes, all the performance studies described (analytical recovery, linearity, method comparison, precision, interference, specificity, stability) represent the standalone performance of the ARK Methotrexate Assay. There is no human-in-the-loop aspect described in these tests; the assay directly provides quantitative results.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc):
The ground truth used for this in vitro diagnostic assay consists of:
- Reference Standards: For accuracy, LoQ, linearity, and specificity, the ground truth was based on precisely known concentrations of methotrexate prepared using certified stock concentrates.
- Predicate Device (Monoclonal FPIA): For method comparison, the ground truth was the results obtained from the legally marketed predicate device (Abbott TDx®/TDxFLx® Methotrexate II), which is an established method for methotrexate measurement.
- Pre-defined Parameters/Conditions: For interference, specificity, and stability studies, the ground truth was the expectation of what the device should measure under specific well-defined conditions (e.g., no cross-reactivity with certain compounds, stability over a given period).
8. The sample size for the training set:
Not applicable. This document describes the validation of an immunoassay kit, not a machine learning or AI algorithm that requires a "training set." Immunoassays are based on biochemical reactions with antibodies, not pattern recognition learned from data.
9. How the ground truth for the training set was established:
Not applicable, as there is no "training set" in the context of this immunoassay device.
Ask a specific question about this device
(169 days)
The ARK™ Gabapentin Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of gabapentin in human serum or plasma on automated clinical chemistry analyzers. Gabapentin concentrations can be used as an aid in management of patients treated with gabapentin.
The ARKTM Gabapentin Calibrator is intended for use in calibration of the ARK Gabapentin Assay.
The ARKTM Gabapentin Control is intended for use in quality control of the ARK Gabapentin Assay.
The ARK Gabapentin Assay is a homogeneous immunoassay based on competition between drug in the specimen and gabapentin labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay.
The ARK Gabapentin Assay consists of reagents R1 anti-gabapentin polyclonal antibody with substrate and R2 gabapentin labeled with bacterial G6PDH enzyme. The ARK Gabapentin Calibrator consists of a six-level set to calibrate the assay, and the ARK Gabapentin Control consists of a three-level set used for quality control of the assay.
ARK Gabapentin products contain ≤0.09% sodium azide. As a precaution, affected plumbing should be flushed adequately with water to mitigate the potential accumulation of explosive metal azides. No special handling is required regarding other assay components.
Here's a summary of the acceptance criteria and study findings for the ARK™ Gabapentin Assay, based on the provided 510(k) summary:
1. Table of Acceptance Criteria and Reported Device Performance
| Performance Metric | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | ≤20% CV with ±15% recovery | 0.75 µg/mL (demonstrated acceptable inter-assay precision and recovery) |
| Recovery / Accuracy | Not explicitly stated as a defined criterion (implied by typical assay validation standards that recovery should be close to 100% within a certain range). | Mean percent recovery: 100.9% across concentrations from 1.0 to 40.0 µg/mL. |
| Linearity | Percent difference between predicted 1st and 2nd order regressed values of ±10%, or ±15% for concentrations ≤ 1.0 µg/mL. | Linear relationship demonstrated between 0.75 and 48.0 µg/mL. All points met the ±10% or ±15% (for lower concentrations) difference criteria. (e.g., 0.75 µg/mL: 12.0% difference; 1.0 µg/mL: 8.4% difference; all others below 2.2% difference up to 48 µg/mL). |
| Assay Range | Not explicitly stated as an acceptance criterion for the range itself, but the device performance defines the reportable range. | 0.75 to 40.0 µg/mL. |
| Method Comparison (Correlation) | Implied by the use of Passing-Bablok regression: a slope close to 1, y-intercept close to 0, and a high correlation coefficient (r²) indicating strong agreement with reference methods. | Study 1 (LC-MS/MS): Slope 0.96 (0.92 to 0.99), y-intercept -0.06 (-0.28 to 0.18), r² 0.96 (0.95 to 0.97). Study 2 (HPLC): Slope 1.08 (1.03 to 1.13), y-intercept -0.08 (-0.35 to 0.25), r² 0.97 (0.95 to 0.98). Study 3 (LC-MS/MS): Slope 1.13 (1.08 to 1.17), y-intercept 0.31 (0.06 to 0.52), r² 0.98 (0.97 to 0.99). |
| Precision | <10% total CV | ARK Gabapentin Control: Low 5.6%, Mid 4.4%, High 3.6% (total CV). Human Serum: Low 7.7%, Mid 4.6%, High 4.7% (total CV). All met the <10% total CV criterion. |
| Interfering Substances | Measurement of gabapentin resulted in ≤10% error in the presence of interfering substances at the levels tested. | All tested substances (Albumin, Bilirubin Conjugated/Unconjugated, Cholesterol, Gamma-Globulin, Hemoglobin, Intralipid®, Rheumatoid Factor, Triglycerides, Uric Acid) resulted in percentage recovery between 95.2% and 106.6% (i.e., within 10% error) for gabapentin concentrations of 2 µg/mL and 20 µg/mL. |
| Drug Interference | Measurement of gabapentin resulted in ≤10% error in the presence of drug compounds at the levels tested. | Most tested anti-epileptic or co-administered drugs and L-amino acids showed percentage recovery within 10% error (between 90% and 110%) for both 2 µg/mL and 20 µg/mL gabapentin. Note: Pregabalin showed higher cross-reactivity at high concentrations (e.g., 100 µg/mL Pregabalin led to 156.9% recovery at 2 µg/mL Gabapentin). The document notes that "Care should be taken when interpreting ARK Gabapentin results if pregabalin is also being administered." |
| Anticoagulants | Not explicitly stated as a numerical criterion, but implies no significant difference. | "The results indicate that there is no significant difference between the recovery of gabapentin in serum or plasma." |
| Sample Stability | Fresh specimens preferred. Clarified specimens: up to one week at 2-8°C. Frozen (≤ -10°C): up to four weeks (acceptance criterion ± 10%). Withstand 3 freeze-thaw cycles. | Specimens shown to meet these criteria. |
| On-Board Stability (Calibration Curve) | Up to 84 days. | Effective up to 84 days based on supporting data. |
| On-Board Stability (Reagent) | Up to at least 84 days. | Effective up to at least 84 days based on supporting data. |
| Accelerated OPEN stability of calibrators and controls | Calibrators and controls stable OPEN at 37℃ for seven (7) days. Once opened: 12 months at 2-8℃. | Shown to be stable per criteria. |
2. Sample Size Used for the Test Set and Data Provenance
- Recovery: Not explicitly stated how many unique samples were tested, but "Six replicates of each sample were assayed." The samples were "human serum negative for gabapentin" spiked with pure gabapentin. Provenance: Not specified (retrospective/prospective, country of origin).
- Linearity: Not explicitly stated how many unique samples, but dilutions were made from a 48.0 µg/mL serum sample. Provenance: Not specified.
- Method Comparison:
- Study 1: 183 samples. Provenance: Not specified.
- Study 2: 64 samples. Provenance: Not specified.
- Study 3: 49 samples. Provenance: Not specified.
- Precision: 160 measurements were taken for each of the three control levels and three human serum pooled specimens (quadruplicate, twice a day for 20 days). Provenance: Not specified.
- Interfering Substances: For each interfering substance, samples were prepared with gabapentin at approximately 2 µg/mL and 20 µg/mL, and assayed against a serum control. The number of individual samples (beyond the spiked ones) isn't specified. Provenance: Not specified.
- Drug Interference: Similar to interfering substances, samples were prepared with gabapentin at approximately 2 µg/mL and 20 µg/mL, spiked with high concentrations of various drugs, and assayed against a serum control. Provenance: Not specified.
- Anticoagulants: "Studies were conducted to determine the performance characteristics of the assay for both serum and plasma samples." Specific sample sizes not provided. Provenance: Not specified.
- Sample Stability: Specific sample sizes not provided, but studies addressed various storage conditions and freeze-thaw cycles. Provenance: Not specified.
It is common for these types of in vitro diagnostic device submissions for a new assay to describe the preparation of test materials and not necessarily the specific provenance of every human sample component (e.g., general "human serum").
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
This type of in vitro diagnostic device (quantitative immunoassay) does not typically involve human experts for establishing "ground truth" in the way an imaging AI device would. Instead, the "ground truth" for the performance studies described is established by:
- Reference methods: For method comparison, reference standards like LC-MS/MS and HPLC are used to define the true concentration of gabapentin in samples. These are highly accurate analytical techniques, not human expert consensus.
- Spiking studies: For recovery, linearity, interference, and drug interference, known quantities of high-purity gabapentin or interfering substances are added to negative serum/plasma, establishing a precise theoretical "ground truth" concentration.
Therefore, the concept of "number of experts" and their "qualifications" for ground truth determination is not applicable in this context.
4. Adjudication Method for the Test Set
Not applicable. As explained above, the ground truth is established by analytical reference methods or known spiked concentrations, not by human interpretation or 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
Not applicable. This device is a quantitative immunoassay for measuring drug concentration in blood, not an imaging device or a diagnostic aid that would involve human "readers" or AI assistance in interpretation.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) was done
Yes, the studies described are all "standalone" in nature, as they assess the performance of the assay itself (the "algorithm only" in a broader sense of an analytical method) to accurately measure gabapentin concentrations. The output of the device is a quantitative value (gabapentin concentration in µg/mL), which is then used by clinicians. There is no "human-in-the-loop performance" component for how the assay itself functions.
7. The Type of Ground Truth Used
The ground truth for the performance studies was established using:
- Reference analytical methods: High Performance Liquid Chromatography - Mass Spectrometry (LC-MS/MS) and High Performance Liquid Chromatography (HPLC) were used as reference methods for the method comparison studies.
- Spiked samples: For recovery, linearity, interference, and drug interference studies, known, precise amounts of pure gabapentin or interfering substances were added to gabapentin-negative human serum/plasma to create samples with known theoretical concentrations.
8. The Sample Size for the Training Set
Not applicable. This is an immunoassay, not a machine learning or AI algorithm in the contemporary sense that would involve a "training set" for model development. The development of the assay (e.g., antibody selection, reagent formulation) is a traditional chemical and biological process, not a computational learning process with distinct training and test sets.
9. How the Ground Truth for the Training Set was Established
Not applicable, as there is no "training set" in the context of this immunoassay.
Ask a specific question about this device
(172 days)
The ARK™ Lamotrigine Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of lamotrigine in human serum or plasma on automated clinical chemistry analyzers. Lamotrigine concentrations can be used as an aid in management of patients treated with lamotrigine.
The ARKTM Lamotrigine Calibrator is intended for use in calibration of the ARK Lamotrigine Assay.
The ARKTM Lamotrigine Control is intended for use in quality control of the ARK Lamotrigine Assay.
The ARK Lamotrigine Assay is a homogeneous immunoassay based on competition between drug in the specimen and lamotrigine labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay.
The ARK Lamotrigine Assay consists of reagents R1 anti-lamotrigine polyclonal antibody with substrate and R2 lamotrigine labeled with bacterial G6PDH enzyme. The ARK Lamotrigine Calibrator consists of a six-level set to calibrate the assay, and the ARK Lamotrigine Control consists of a three-level set used for quality control of the assay.
The ARK™ Lamotrigine Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of lamotrigine in human serum or plasma. The device performance was evaluated through various studies, including limit of quantitation (LOQ), assay range, recovery, linearity, method comparison, precision, interfering substances, specificity, and anticoagulant studies.
1. Acceptance Criteria and Reported Device Performance:
| Study/Characteristic | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | <20% CV with ±15% recovery | 0.85 µg/mL |
| Assay Range | Not explicitly stated as acceptance criteria, but defined as the operational range. | 0.85 to 40.00 µg/mL |
| Recovery | Not explicitly stated as acceptance criteria, but individual recovery values are presented. | Mean percent recovery: 99.2%. Individual recoveries ranged from 95.8% to 105.1%. |
| Linearity | Percent difference ±10% between predicted and 2nd order regressed values. | All tested concentrations showed a percentage difference within ±10% (range -0.4% to 7.1%). |
| Method Comparison (Study 1 - HPLC) | Not explicitly stated as acceptance criteria, but presented with 95% confidence limits. | Slope: 1.01 (0.99 to 1.03), y-intercept: 0.37 (0.22 to 0.55), Correlation Coefficient (r²): 0.97 (0.96 to 0.98) |
| Method Comparison (Study 2 - Turbidimetric Immunoassay) | Not explicitly stated as acceptance criteria, but presented with 95% confidence limits. | Slope: 0.93 (0.89 to 0.97), y-intercept: 0.41 (0.07 to 0.74), Correlation Coefficient (r²): 0.96 (0.94 to 0.97) |
| Precision | ≤10% total CV | All tested samples (low, mid, high controls, and human serum pools) showed total CVs well within 10% (range 4.1% to 8.8%). |
| Interfering Substances | ≤10% error in the presence of interfering substances. | Measurement of lamotrigine resulted in ≤10% error in the presence of various interfering substances (e.g., albumin, bilirubin, cholesterol, hemoglobin, etc.) at tested levels. |
| Specificity (Metabolites) | Not explicitly stated as acceptance criteria for cross-reactivity percentage. | Metabolites showed low cross-reactivity: Lamotrigine-2-N-glucuronide (1.09-2.91%), Lamotrigine-2-N-methyl (0.02-0.24%), Lamotrigine-2-N-oxide (1.30-3.94%). |
| Specificity (Drug Interference) | ≤10% error in the presence of co-administered drugs. | Measurement of lamotrigine resulted in ≤10% error in the presence of a wide range of co-administered drugs at tested levels. |
| Cross-Reacting Drug (Trimethoprim) | Caution advised if trimethoprim is administered, no explicit error threshold for this specific drug. | Trimethoprim at 40.0 µg/mL showed 4.4% cross-reactivity at 3 µg/mL lamotrigine and 3.0% cross-reactivity at 15 µg/mL lamotrigine. Percentage recovery was 156.0% (at 3 µg/mL Lamotrigine) and 108.0% (at 15 µg/mL Lamotrigine), indicating significant interference. |
| Anticoagulants | No significant difference between serum and plasma recovery. | Results indicated no significant difference between the recovery of lamotrigine in serum or plasma. |
| Sample Stability | Defined stability periods. | Serum specimens stable for at least 6 months frozen, 50 hours at room temperature (22°C), 37 days refrigerated (2-8°C), and after 3 freeze/thaw cycles. |
| On-Board Stability (Calibration Curve) | Calibration curve stability for 30 days. | Calibration curve stability for a period of 30 days is supported by data. |
| On-Board Stability (Reagent) | Reagents effective for up to 30 days. | Reagents were effective when stored on-board for up to at least 30 days. |
| On-Board Stability (Calibrator/Controls) | In-use stability demonstrated; 12 months opened at 2-8°C after accelerated stability. | In-use stability of calibrators and controls was demonstrated. Accelerated OPEN stability at 37°C for 7 days showed stability. Once opened, vials may be stored at 2-8°C for 12 months. |
2. Sample Size and Data Provenance for Test Set:
- Recovery: Not explicitly stated, but "six replicates of each sample" were assayed across various concentrations, implying a total of 48 individual assays. Data provenance is human serum negative for lamotrigine, spiked with concentrated drug.
- Linearity: Not explicitly stated, but dilutions from a 48.00 µg/mL serum sample were made. Data provenance is human serum negative for lamotrigine, spiked with drug.
- Method Comparison (Study 1 - HPLC): 193 samples. Provenance not explicitly stated (e.g., country of origin, retrospective/prospective), but implied to be clinical samples or samples prepared to simulate clinical range.
- Method Comparison (Study 2 - Turbidimetric Immunoassay): 77 samples. Provenance not explicitly stated.
- Precision: Tri-level controls and three human serum pooled specimens. Each level assayed in quadruplicate twice a day for 20 days (160 measurements per control/pool level).
- Interfering Substances: Clinically high concentrations of substances in human serum with known lamotrigine levels (approximately 3 and 15 µg/mL).
- Specificity (Metabolites): Metabolites spiked into two separate samples each containing low (3 µg/mL) and high (15 µg/mL) therapeutic levels of lamotrigine.
- Specificity (Drug Interference): High concentration of each compound spiked into normal human serum with known lamotrigine levels (approximately 3 and 15 µg/mL).
- Anticoagulants: Not explicitly stated but implies a set of serum and plasma samples were tested.
- Sample Stability: Not explicitly stated, but samples were subject to various conditions (frozen, room temp, refrigerated, freeze/thaw cycles).
3. Number of Experts and their Qualifications for Ground Truth:
This type of device (quantitative immunoassay) relies on analytical performance against established reference methods or known concentrations, rather than expert interpretation of images or clinical data. Therefore, the concept of "experts establishing ground truth" as it applies to subjective assessments is not directly applicable here. The ground truth for analytical studies like recovery and linearity are based on gravimetric dilutions/known concentrations, and for method comparisons, the reference methods themselves (HPLC, predicate immunoassay) serve as the comparative standard.
4. Adjudication Method for the Test Set:
Not applicable for this type of analytical device validation. The studies involve quantitative measurements and comparisons to reference methods or known values, not subjective interpretations requiring adjudication.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
No, an MRMC comparative effectiveness study was not done. This type of study is typically performed for diagnostic devices requiring human interpretation of results, such as imaging studies, where the AI's effect on human reader performance is evaluated. This device is an automated immunoassay for quantitative measurement.
6. Standalone (Algorithm Only) Performance:
Yes, the entire submission describes the standalone performance of the ARK™ Lamotrigine Assay. The studies (LOQ, assay range, recovery, linearity, method comparison, precision, interference, specificity, stability) evaluate the device's analytical performance independently. The device operates on automated clinical chemistry analyzers without human intervention for result generation.
7. Type of Ground Truth Used:
- Known Concentrations/Gravimetric Dilutions: For studies such as Limit of Quantitation, Recovery, Linearity, Precision, Interfering Substances, Specificity (Metabolites, Drug Interference), and Anticoagulants, the ground truth is established by preparing samples with known, precise concentrations of lamotrigine or interfering substances through gravimetric dilutions.
- Reference Methods: For method comparison studies, the reference methods are High Performance Liquid Chromatography (HPLC) and a predicate turbidimetric immunoassay (QMS® Lamotrigine, K062966).
8. Sample Size for the Training Set:
Not applicable. This device is a quantitative immunoassay with "wet lab" validation, not an AI/ML algorithm that typically requires a large clinical "training set" to learn patterns. The validation process involves demonstrating analytical performance against known standards and comparative methods.
9. How Ground Truth for the Training Set Was Established:
Not applicable, as there isn't a "training set" in the context of an AI/ML algorithm. The device's analytical characteristics are determined through laboratory experiments using carefully prepared samples with known concentrations or by comparison to reference methods.
Ask a specific question about this device
(168 days)
The ARK™ Zonisamide Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of zonisamide in human serum or plasma samples on automated clinical chemistry analyzers. Zonisamide concentrations can be used as an aid in management of patients treated with zonisamide.
The ARK™ Zonisamide Calibrator is intended for use in calibration of the ARK Zonisamide Assay.
The ARK™ Zonisamide Control is intended for use in quality control of the ARK Zonisamide Assav.
The ARK Zonisamide Assay is a homogeneous immunoassay based on competition between drug in the specimen and zonisamide labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay.
The ARK Zonisamide Assay consists of reagents R1 anti-zonisamide polyclonal antibody with substrate and R2 zonisamide labeled with bacterial G6PDH enzyme. The ARK Zonisamide Calibrator consists of a six-level set to calibrate the assay, and the ARK Zonisamide Control consists of a three-level set used for quality control of the assay.
Here's a breakdown of the acceptance criteria and the study details for the ARK Zonisamide Assay, based on the provided 510(k) summary:
Acceptance Criteria and Device Performance
| Test/Characteristic | Acceptance Criteria (Implicit) | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | <20% CV with ±15% recovery | 2.0 µg/mL (with corresponding recovery and precision) |
| Recovery | % Recovery close to 100% (within reasonable analytical limits, e.g., ±15%) | Ranges from 85.3% to 110.0% for concentrations 2.0 - 50.0 µg/mL |
| Linearity | % Difference ±10% between predicted 1st and 2nd order regressed values (±15% below 3.0 µg/mL) | Ranges from -7.0% to 2.7% for concentrations 2.4 - 48.0 µg/mL |
| Assay Range | Defined range of quantitative measurement | 2.0 to 50.0 µg/mL |
| Method Comparison (Correlation Coefficient) | High correlation (e.g., r² > 0.90) with predicate device | r² = 0.93 (0.91 to 0.95 95% CI) |
| Precision (Total CV) | <10% Total CV | Ranges from 4.5% to 5.9% for various concentration levels |
| Interfering Substances | Error ≤10% in the presence of listed interferents | Error ≤10% for all tested interfering substances |
| Drug Interference | Error ≤10% in the presence of listed drug compounds | Error ≤10% for all tested drug compounds |
| Anticoagulants | No significant difference in recovery between serum and plasma | No significant difference between serum and plasma samples |
| Sample Stability (Room Temp) | Stable for at least 24 hours | Stable for at least 24 hours (22°C) |
| Sample Stability (Refrigerated) | Stable for a specified duration | Stable for 28 days (2-8°C) |
| Sample Stability (Frozen) | Stable for a specified duration | Stable for 56 days |
| Sample Stability (Freeze/Thaw) | Stable after 3 successive freeze/thaw cycles | Stable after 3 successive freeze/thaw cycles |
| Calibration Curve Stability | Effective for a specified duration | Effective up to 46 days |
| Reagent On-Board Stability | Effective for a specified duration | Effective up to 32 days (uncapped) and 46 days (capped) |
Study Details
-
Sample sizes used for the test set and the data provenance:
- LOQ: 20 replicates for each sample. Data provenance not specified (likely laboratory-prepared samples).
- Recovery: 20 replicates for each sample. Data provenance not specified (likely laboratory-prepared samples with added drug).
- Linearity: Zonisamide concentrations ranged from 0.8 to 80.0 ug/mL, with dilutions made proportionally. Data provenance not specified (likely laboratory-prepared samples).
- Method Comparison: 176 samples. Data provenance not specified.
- Precision: 160 replicates per control level and human serum level (quadruplicate twice a day for 20 days). Data provenance not specified (likely laboratory-prepared controls and pooled human serum specimens).
- Interfering Substances: Zonisamide levels of approximately 15 and 45 µg/mL, spiked with various interferents. Specific number of samples per interferent not specified, but each was "evaluated." Data provenance not specified (likely laboratory-prepared samples).
- Metabolites: Zonisamide levels of 15 µg/mL and 45 µg/mL, spiked with NAZ (50.0, 10.0 µg/mL) and SMAP (50.0, 10.0 µg/mL). Data provenance not specified (likely laboratory-prepared samples).
- Drug Interference: Zonisamide levels of approximately 15 and 45 µg/mL, spiked with various drug compounds. Specific number of samples per drug not specified, but each was "assayed." Data provenance not specified (likely laboratory-prepared samples).
- Anticoagulants: Not explicitly stated, but "studies were conducted" on serum and plasma. Data provenance not specified.
- Sample Stability: Not explicitly stated, but "serum specimens were shown to be stable" under various conditions. Data provenance not specified.
- On-Board Stability: "Supporting data" cited, but specific sample sizes and provenance not detailed.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- No expert ground truth was described for the test set. The ground truth for analytical performance studies (LOQ, recovery, linearity, precision) is typically established by precisely preparing samples with known concentrations. For method comparison, the predicate device serves as the comparative "truth."
-
Adjudication method for the test set:
- Not applicable, as this is an in vitro diagnostic assay and not an image-based or qualitative diagnostic device requiring expert adjudication.
-
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 assay, not an AI-assisted diagnostic tool for human readers.
-
If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
- Yes, the entire submission describes the standalone performance of the ARK Zonisamide Assay. Its function is to quantitatively determine zonisamide concentrations in samples, which is an algorithm-only (analytical assay) performance.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
- Analytical Ground Truth: For most studies (LOQ, recovery, linearity, interfering substances, drug interference, metabolites), the ground truth was established through gravimetrical or volumetric preparation of samples with known, spiked concentrations of zonisamide or interfering substances.
- Comparative Ground Truth: For the method comparison study, the predicate device (QMS® Zonisamide Assay) served as the reference for comparison.
-
The sample size for the training set:
- The document does not describe a "training set" in the context of machine learning or AI. This assay is a chemical immunoassay, not a learning algorithm. The studies described are analytical validation studies to characterize the performance of the chemical reaction and measurement system.
-
How the ground truth for the training set was established:
- As there is no described training set, this question is not applicable.
Ask a specific question about this device
(146 days)
The ARK™ Levetiracetam Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of levetiracetam in human serum or plasma on automated clinical chemistry analyzers. Levetiracetam concentrations can be used as an aid in management of patients treated with levetiracetam. The ARKTM Levetiracetam Calibrator is intended for use in calibration of the ARK Levetiracetam Assay. The ARK™ Levetiracetam Control is intended for use in quality control of the ARK Levetiracetam Assay.
The ARK Levetiracetam Assay is a homogeneous immunoassay based on competition between drug in the specimen and levetiracetam labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay. The ARK Levetiracetam Assay consists of reagents RI anti-levetiracetam polyclonal antibody with substrate and R2 levetiracetam labeled with bacterial G6PDH enzyme. The ARK Levetiracetam Calibrator consists of a six-level set to calibrate the assay, and the ARK Levetiracetam Control consists of a three-level set used for quality control of the assay.
Here's a summary of the acceptance criteria and the study that proves the device meets them, based on the provided text:
Acceptance Criteria and Device Performance for ARK™ Levetiracetam Assay
1. Table of Acceptance Criteria and Reported Device Performance
| Performance Characteristic | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | 20% CV with ±15% recovery | 2.0 µg/mL (not explicitly stated if it met the 20% CV & ±15% recovery, but implied as the determined LOQ) |
| Accuracy (Analytical Recovery) | Not explicitly stated, but implied to be acceptable based on percent recovery within reasonable limits. | Range of 94.6% to 105.3% recovery for concentrations 2.0 to 100.0 µg/mL. |
| Linearity | Percent difference ±10% between 1st and 2nd order regressed values, or ±15% below 3.0 µg/mL. | Linear relationship demonstrated between 2.0 and 100.0 µg/mL. All % Differences were within the specified ±10% (for values ≥3.0 µg/mL) and ±15% (for values <3.0 µg/mL, observed at 13.2% for 2.0 µg/mL which is within 15%). |
| Precision | <10% total CV | ARK Levetiracetam Control:LOW (7.5 µg/mL): 4.5% total CVMID (29.4 µg/mL): 3.7% total CVHIGH (73.4 µg/mL): 4.2% total CVHuman Serum:LOW (6.9 µg/mL): 4.8% total CVMID (30.2 µg/mL): 4.1% total CVHIGH (75.5 µg/mL): 4.4% total CV |
| Interfering Substances | Measurement of levetiracetam resulted in ≤10% error. | Measurement of levetiracetam resulted in ≤10% error in the presence of tested interfering substances (Albumin, Bilirubin, Cholesterol, Gamma-Globulin, Hemoglobin, Intralipid®, Rheumatoid Factor, Triglycerides, Uric Acid). |
| Metabolite Cross-Reactivity (ucb L057) | Measurement of levetiracetam resulted in ≤10% error. | Measurement of levetiracetam resulted in ≤10% error. Specifically, percent interference was -3.0% (at 15 µg/mL Levetiracetam) and 6.6% (at 50 µg/mL Levetiracetam). Percent cross-reactivity was -0.2% and 1.3%. |
| Drug Interference | Measurement of levetiracetam resulted in ≤10% error. | Measurement of levetiracetam resulted in ≤10% error in the presence of various anti-epileptic or co-administered drugs tested. All percentage recoveries listed for 15 µg/mL and 50 µg/mL Levetiracetam were within 10% of the expected value (i.e., between 90% and 110%). |
| Anticoagulants | No significant difference between recovery in serum or plasma. | Results indicate no significant difference between the recovery of levetiracetam in serum or plasma. |
| Sample Stability | Not explicitly stated beyond "stable for at least...". | Stable for at least 48 hours at room temperature (22 °C), 40 days when refrigerated (2-8 °C), and after three successive freeze/thaw cycles. |
| Calibration Curve Stability | Not explicitly stated beyond "effective up to...". | Effective up to 40 days. |
| Reagent On-Board Stability | Not explicitly stated beyond "effective for up to at least...". | Effective for up to at least 40 days. |
2. Sample Size and Data Provenance
- Accuracy: Not explicitly stated, but "highly pure levetiracetam was added volumetrically to human serum negative for levetiracetam, representing drug concentrations across the assay range." Six replicates of each sample were assayed. Data provenance not specified (retrospective/prospective, country of origin).
- Linearity: Not explicitly stated, but a 100.0 µg/mL serum sample was prepared and dilutions were made proportionally with human serum negative for levetiracetam. Data provenance not specified.
- Method Comparison: Number of samples = 305. Data provenance not specified.
- Precision: 160 measurements for each of the 6 levels (3 control levels, 3 human serum pools). This equates to 8 measurements per level per day (quadruplicate twice a day) over 20 days. Data provenance not specified.
- Interfering Substances: For each interfering substance, serum with known levels of levetiracetam (approx. 15 and 50 µg/mL) was evaluated. A serum control was also used. Amount of samples for each interfering substance is not explicitly stated. Data provenance not specified.
- Metabolites: Not explicitly stated for metabolite cross-reactivity. Data provenance not specified.
- Drug Interference: For each compound, normal human serum with known levels of levetiracetam (approx. 15 and 50 µg/mL) was spiked and assayed along with a serum control. Total number of samples is not explicitly stated. Data provenance not specified.
- Anticoagulants: Not explicitly stated. Data provenance not specified.
3. Number of Experts and Qualifications for Ground Truth
Not applicable. This is an in-vitro diagnostic device (immunoassay) for quantitative determination of a drug. Ground truth is established by spiked concentrations, a reference method (LC/MS/MS), or intrinsic characteristics of known substances. No human expert interpretation of images or other subjective data is involved.
4. Adjudication Method
Not applicable. For an immunoassay, the "ground truth" is typically verifiable concentrations of the analyte or comparison to a gold standard analytical method. There is no subjective assessment that would require expert adjudication.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No. This is an in-vitro diagnostic device for quantitative chemical analysis, not an imaging device or a device involving human interpretation where MRMC studies would be relevant.
6. Standalone Performance
Yes, the studies described are solely for the performance of the ARK™ Levetiracetam Assay system (algorithm/device only). There is no "human-in-the-loop" component to its stated intended use or the performance studies presented.
7. Type of Ground Truth Used
- Limit of Quantitation, Accuracy, Linearity, Interfering Substances, Metabolites, Drug Interference: Ground truth was established by known, spiked concentrations of levetiracetam and/or interfering substances into human serum, or by using highly pure reference materials.
- Method Comparison: Ground truth for comparison was established by a reference LC/MS/MS method.
- Precision: Ground truth was established by known concentrations in control materials and pooled human serum samples.
8. Sample Size for the Training Set
Not applicable. This is an immunoassay, which does not typically involve training a machine learning model in the same way as, for example, an AI for image analysis. The "training" in this context would be the development and optimization of the chemical reagents and assay parameters during the R&D phase, for which specific sample sizes and datasets for "training" are not usually reported in this manner.
9. How the Ground Truth for the Training Set Was Established
Not applicable. As noted above, this is not a machine learning device that uses a "training set" in the conventional sense. The "ground truth" for the development of the assay would involve fundamental chemical and biological principles to ensure proper antibody-antigen binding specificity and enzyme activity, calibrated against known standards.
Ask a specific question about this device
(115 days)
The ARK™ Topiramate Assay is a homogeneous enzyme immunoassay intended for the quantitative determination of topiramate in human serum or plasma on automated clinical chemistry analyzers. The results obtained are used in the diagnosis and treatment of topiramate overdose and in monitoring levels of topiramate to help ensure appropriate therapy.
The ARK™ Topiramate Calibrator is intended for use in calibration of the ARK Topiramate Assav.
The ARKTM Topiramate Control is intended for use in quality control of the ARK Topiramate Assay.
The ARK Topiramate Assay is a homogeneous immunoassay based on competition between drug in the specimen and topiramate epitope labeled with the enzyme glucose-6-phosphate dehydrogenase (G6PDH) for binding to the antibody reagent. As the latter binds antibody, enzyme activity decreases. In the presence of drug from the specimen, enzyme activity increases and is directly proportional to the drug concentration. Active enzyme converts the coenzyme nicotinamide adenine dinucleotide (NAD) to NADH that is measured spectrophotometrically as a rate of change in absorbance. Endogenous serum G6PDH does not interfere with the results because the coenyzme NAD functions only with the bacterial enzyme used in the assay.
The ARK Topiramate Assay consists of reagents R1 anti-topiramate polyclonal antibody with substrate and R2 topiramate epitope labeled with bacterial G6PDH enzyme. The ARK Topiramate Calibrator consists of a six-level set to calibrate the assay, and the ARK Topiramate Control consists of a three-level set used for quality control of the assay.
Here's a breakdown of the acceptance criteria and the studies that prove the device meets them, based on the provided 510(k) summary:
1. Table of Acceptance Criteria and Reported Device Performance
| Performance Characteristic | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Limit of Quantitation (LOQ) | ≤20% CV with ±15% recovery | 1.5 µg/mL |
| Accuracy (Analytical Recovery) | Percent recoveries within 10% of theoretical levels | All reported values from 1.5 µg/mL to 55.0 µg/mL were within 10% of theoretical (e.g., 95.6% to 107.1% recovery). Example: 1.5 µg/mL recovered 95.6%, 55.0 µg/mL recovered 107.1%. |
| Linearity | Percent difference ≤ ±10% between predicted 1st and 2nd order regressed values | Demonstrated linearity between 1.2 and 54.0 µg/mL. Max % Difference outside this range was -18.14% at 0.6 µg/mL. All values within 1.2-54.0 µg/mL were within ±10%. |
| Assay Range | Not explicitly stated as acceptance criteria, but defined as output. | 1.5 µg/mL to 54.0 µg/mL |
| Precision (Total CV) | <10% total CV | Sample 1 (2.4 µg/mL): 4.2% CVSample 2 (10.2 µg/mL): 2.7% CVSample 3 (40.2 µg/mL): 3.2% CV |
| Interfering Substances | Measurement of topiramate resulted in ≤10% error | All tested substances (e.g., Albumin, Bilirubin, Cholesterol, Hemoglobin) at clinically high concentrations resulted in ≤10% error. |
| Metabolites (Cross-Reactivity) | Measurement of topiramate resulted in ≤10% error | 9-Hydroxy-topiramate at 40.0 µg/mL resulted in 8.6% error for low topiramate and 3.2% error for high topiramate. |
| Drug Interference | Measurement of topiramate resulted in ≤10% error | All tested co-administered drugs (e.g., Acetaminophen, Carbamazepine, Phenobarbital, Valproic Acid) at high concentrations resulted in ≤10% error. |
| Anticoagulants | No significant difference between serum and plasma recovery | Results indicated no significant difference between serum and plasma recovery. |
| Calibration Curve Stability | Not explicitly stated as acceptance criteria, but reported. | Effective up to 49 days. |
| Reagent On-board Stability | Not explicitly stated as acceptance criteria, but reported. | Effective for up to 60 days. |
2. Sample Size Used for the Test Set and Data Provenance
- Accuracy: Not explicitly stated as a "test set" in the traditional sense. Samples were human serum negative for topiramate, spiked with known topiramate concentrations. Six replicates were performed for each concentration. Data provenance is implied to be laboratory-generated through spiking.
- Linearity: A 60.0 µg/mL serum sample was prepared and proportionally diluted with human serum negative for topiramate. Topiramate concentrations ranged from 0.6 to 60.0 µg/mL. Data provenance is laboratory-generated.
- Method Comparison: 113 samples.
- Data Provenance: Not explicitly stated for this particular study (e.g., country of origin, retrospective/prospective). However, given the context of a 510(k) for an in vitro diagnostic, these samples would typically come from patients, likely collected retrospectively or prospectively for method validation purposes.
- Precision: Tri-level controls were used. Each level was assayed in quadruplicate, twice a day for 20 days (N=160 for each level). Data provenance is laboratory-generated.
- Interfering Substances: Samples were human serum with known levels of topiramate (approximately 5 and 20 µg/mL) spiked with high concentrations of various interfering substances. Data provenance is laboratory-generated.
- Metabolites: Samples were human serum with known levels of topiramate spiked with 9-hydroxy-topiramate. Data provenance is laboratory-generated.
- Drug Interference: Samples were normal human serum with known levels of topiramate (approximately 5 and 20 µg/mL) spiked with high concentrations of various drug compounds. Data provenance is laboratory-generated.
- Anticoagulants: Not explicitly detailed, but implied to involve testing serum and plasma samples containing topiramate. Data provenance is laboratory-generated.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
Not applicable. This is an in vitro diagnostic device for quantitative determination of a drug in serum/plasma. The "ground truth" for the test set (e.g., known concentrations for accuracy, reference method results for method comparison) is established through analytical techniques, not expert adjudication of images or clinical cases.
4. Adjudication Method for the Test Set
Not applicable. As noted above, this device performance is assessed through analytical measurements against known concentrations or a reference method, not through expert adjudication in the context of diagnostic interpretation.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done
No, an MRMC comparative effectiveness study was not done. This type of study is typically conducted for diagnostic imaging devices where human readers interpret cases, and the AI's impact on their performance is evaluated. This device is an automated in vitro diagnostic assay.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
Yes, the studies presented are all standalone validations of the analytical performance of the ARK Topiramate Assay. The device is intended for quantitative determination on automated clinical chemistry analyzers, meaning it operates without direct human-in-the-loop interpretation of its primary output.
7. The Type of Ground Truth Used
The ground truth used for these studies includes:
- Known Reference Concentrations: For accuracy and linearity studies, serum samples were spiked with precisely measured, highly pure topiramate to create samples with known theoretical concentrations.
- Reference Method Results: For the method comparison study, results from the ARK Topiramate Assay were compared against a "commercially available FPIA Immunoassay." The results from this predicate method served as the reference or ground truth for comparison.
- Known Spiked Levels: For interference, metabolite, and drug interference studies, known quantities of the interfering substance, metabolite, or drug were added to samples with known topiramate concentrations.
8. The Sample Size for the Training Set
Not applicable. This document describes the validation of a predefined assay, not the development or training of a machine learning algorithm. Therefore, there is no "training set" in the context of an AI/ML device. The reagents are pre-formulated for the immunoassay.
9. How the Ground Truth for the Training Set was Established
Not applicable for the same reason as point 8.
Ask a specific question about this device
Page 1 of 1