Search Results

The Focus Diagnostics' STRATIFY JCV® DxSelect™ assay is intended for the qualitative detection of antibodies to JC virus in human serum or plasma. The assay is intended for use in conjunction with other clinical data, in multiple sclerosis patients receiving or considering natalizumab therapy, as an aid in risk stratification for progressive multifocal leukoencephalopathy development. The assay is for professional use only.

The assay is not intended for donor screening. The performance of this assay has not been established for use in other immunocompromised patient populations or patients with different disease conditions or undergoing other treatments or in neonates and pediatric patient populations.

The Focus Diagnostics' STRATIFY JCV® DxSelect™ test is an ELISA assay. JC virus-like particles (VLP) are pre-coated onto 96-well microiter plates. Diluted serum or plasma specimens and controls are incubated in the wells to allow JCV-specific antibodies present in the specimens to react with the JC VLP antigen. Nonspecific reactants are removed by washing. Peroxidase-conjugated anti-human antibodies are added to react with JCV-specific antibodies. Excess conjugate is removed by washing. Enzyme substrate and chromogen are added, and the color is allowed to develop. After adding the Stop Reagent, the resultant color change is quantified by a spectrophotometric reading of optical density (OD). Specimen OD readings are compared with Cut-Off Calibrator OD readings to determine results. Each specimen result is reported as an index value. A specimen with an index value that is greater than a specified upper cut-off is reported as positive for detectable JCV-specific antibodies, whereas a specimen with an index value less than the specified lower cut-off is reported as negative for detectable JCV-specific antibodies. A specimen with an index value that is equal to or between the upper and lower cut-off values is reported as indeterminate. An indeterminate result requires further evaluation in the confirmation (inhibition) assay.

In the confirmation assay, soluble JC VLP antigen will compete with plate bound JC VLP antigen for the JCV-specific antibodies present in the serum or plasma specimens. After washing away the unbound antibodies, peroxidase-conjugated anti-human antibodies are added and bind to any captured JCVspecific antibodies. Excess conjugate is removed by washing. Enzyme substrate and chromagen are added, and the color is allowed to develop. After adding the Stop Reagent, the resultant color change is quantified by a spectrophotometric reading of OD. The percent inhibition is calculated to confirm presence of JCV-specific antibodies in the specimens with a percent inhibition value that is greater than the specified cut-off are reported as positive for detectable JCV-specific antibodies, whereas specimens with percent inhibition values less than or equal to the cut-off are reported as negative for detectable JCV-specific antibodies.

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:

Acceptance Criteria and Device Performance for STRATIFY JCV® DxSelect™

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a "table of acceptance criteria" in the traditional sense, but rather demonstrates performance against the predicate device and established clinical guidelines. The key performance metrics evaluated are Positive Percent Agreement (PPA) and Negative Percent Agreement (NPA) with a validated laboratory methodology (the predicate device) and JCV antibody positivity rate in certain patient populations.

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

• PML Risk Stratification / Pre-PML samples:
  • Sample Size: 31 available serum samples from confirmed PML patients.
  • Data Provenance: Data collected from both clinical trial and post-marketing reports of confirmed cases of PML (retrospective, likely international given the nature of clinical trials for natalizumab, but not explicitly stated). Samples were collected at least 6 months prior to clinical diagnosis of PML.
• Archived Clinical Specimens (Comparison to Predicate):
  • Sample Size:
    • Patients Receiving Natalizumab: 707 samples (resulted in 414 positive, 293 negative, for a total of 707 evaluable samples based on the table).
    • Patients Considering Natalizumab: 623 samples (resulted in 350 positive, 273 negative, for a total of 623 evaluable samples based on the table).
  • Data Provenance: Prospectively collected and archived clinical samples obtained from the STRATIFY-2 and AFFIRM clinical studies. The AFFIRM study included patients from "North America and EU/Rest of World," while STRATIFY-2 was conducted in the "US." This indicates a multi-national, prospective collection.
• Reproducibility:
  • Sample Size: A panel consisting of four levels of sera and four levels of plasma (EDTA) samples, plus low and high positive controls. Each panel member was tested in three replicates, two runs a day for five days. The total number of quantitative results for Controls and plasma/serum samples listed in Table 5 is typically 90 or 120 per specific parameter, representing replicates across conditions.
  • Data Provenance: Not explicitly stated, but it describes a laboratory study, likely an internal development/validation study by Focus Diagnostics.
• Reproducibility at the Lower Cut Point:
  • Sample Size: One contrived serum sample and one contrived plasma (EDTA) sample. Each was diluted 20 times and tested. (N=20 for plasma, N=19 for serum in the table, likely referring to the number of replicates after dilution).
  • Data Provenance: Laboratory study, likely internal.
• Cross-Reactivity (Part One - Spiked Antibodies):
  • Sample Size: 3 replicates for each of 3 antibodies (total of 9 tests).
  • Data Provenance: Laboratory study.
• Cross-Reactivity (Part Two - Sero-prevalence Comparison):
  • Sample Size: Remnant specimens for 13 different potential cross-reactants, with sample sizes ranging from 17 to 48 (Total 334 samples).
  • Data Provenance: Not explicitly stated, but these are "remnant specimens that previously tested positive for each potential cross reacting antibody," implying clinical samples.
• Cross-Reactivity (Part Three - Polyoma Viruses):
  • Sample Size: 40 clinical specimens previously tested positive for JCV antibodies.
  • Data Provenance: Clinical samples.
• Interferences:
  • Sample Size: Sera and plasma samples containing JCV antibody near the cut-off. Each interferent was tested (N not explicitly stated for each, but inferred from the "Average Index" values).
  • Data Provenance: Laboratory study.
• Hook Effect:
  • Sample Size: A panel of serum samples "ranging from negative reactivity to strong positive reactivity" (not a specific number given).
  • Data Provenance: Laboratory study.
• Sample Matrix Comparison:
  • Sample Size: 109 paired sera and plasma samples.
  • Data Provenance: Not explicitly stated, but likely clinical samples.
• Sample Comparison - Fresh vs. Frozen:
  • Sample Size: 53 pairs of fresh and frozen plasma specimens and 53 pairs of fresh and frozen sera specimens (total 106 pairs).
  • Data Provenance: Not explicitly stated, but likely clinical samples.

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

The document does not mention the use of experts to establish the ground truth for the test set.

• For the comparative agreement studies, the "ground truth" for the test set (pre-PML samples and archived clinical specimens) was established by comparison to a "validated laboratory methodology" (the predicate device, STRATIFY JCV™ Antibody ELISA).
• For PML risk stratification, clinical diagnosis of PML based on existing medical criteria was used as the outcome for the "pre-PML" samples.
• The "control" samples for reproducibility and cross-reactivity studies were internally prepared or sourced as part of the laboratory validation process.

4. Adjudication Method for the Test Set

The document does not describe an adjudication method involving multiple readers for the test set. The comparison studies rely on the results generated by the predicate device as the reference, which presumably has its own established interpretation and quality control.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and Effect Size of Human Improvement with AI vs. Without AI Assistance

No MRMC comparative effectiveness study was done. This device is an in vitro diagnostic (IVD) assay (ELISA-based), not an AI-assisted diagnostic imaging or interpretation tool. Therefore, the concept of "human readers improve with AI vs. without AI assistance" does not apply to this specific device.

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

The device is an immunoassay (ELISA) intended for "professional use only". It produces quantitative results (OD, Index values, %Inhibition) that are then interpreted against established cut-offs to yield qualitative results (Positive, Negative, Indeterminate). While the interpretation logic (cut-offs) is algorithmic, the test requires human involvement for sample preparation, running the assay, and reading the results from a spectrophotometer. The final interpretation of the test result (especially for indeterminate results requiring confirmation) and its use in risk stratification in conjunction with other clinical data explicitly involves a human healthcare professional. Therefore, it's not a standalone "algorithm only" device in the sense of a fully automated AI system rendering a diagnosis.

7. The Type of Ground Truth Used

The ground truth used varies based on the type of study:

• Clinical Performance (PML Risk Stratification):
  • For pre-PML samples: Clinical diagnosis of PML (confirmed cases) combined with the patient's JCV antibody status by the predicate device.
  • For risk estimation: Outcomes data (incidence of PML) from natalizumab clinical trials and post-marketing reports, stratified by JCV serological status.
• Comparative Agreement Studies (Archived Clinical Specimens):
  • Results from the predicate device (STRATIFY JCV™ Antibody ELISA), which is a validated laboratory methodology.
• Reproducibility, Cross-Reactivity, Interferences, Hook Effect, Sample Matrix/Fresh vs. Frozen:
  • Reference materials, contrived samples, or previously characterized clinical samples with known (or expected) JCV antibody status or presence of interfering/cross-reacting substances. These are essentially "expert consensus" or "reference method" results for analytical performance.

8. The Sample Size for the Training Set

The document does not explicitly describe a "training set" in the context of machine learning. As this is an ELISA assay, the development and optimization process would involve analytical studies to determine optimal reagent concentrations, incubation times, and optical density (OD) readings. The cut-off values (Index < 0.2, negative; Index ≥ 0.4, positive; 0.2-0.4 indeterminate) and Confirmation Assay cut-offs are likely established through extensive analytical validation and clinical correlation studies, but these are not explicitly termed "training set" data.

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

As noted above, there is no explicit "training set" described as such. The establishment of cut-off values and assay parameters would typically be done through:

• Extensive testing of characterized clinical samples: This involves samples with known JCV antibody status (determined by a gold standard or comprehensive clinical assessment), alongside samples representing different levels of antibody reactivity.
• Statistical analysis: To define thresholds that optimize sensitivity and specificity.
• Correlation with clinical outcomes: As demonstrated by the PML risk stratification data, where the assay's results correlated with the incidence of PML, thus validating the clinical relevance of the chosen cut-offs.

The predicate device and its established performance also served as a critical benchmark in defining the desired performance characteristics during the development of STRATIFY JCV® DxSelect™.

Ask a specific question about this device