(233 days)
The RapidFRET Oral Fluid Assay for PCP is a homogeneous time-resolved fluorescence assay that is intended for prescription use in central laboratories only on the RapidFRET Integrated Workstation. The assay is used to perform a qualitative screen for Phencyclidine at 10 ng/mL in neat oral fluid samples collected with the RapidEASE Oral Fluid Collector. This assay provides only a preliminary result. To obtain a confirmed analytical result, a more specific alternate chemical method such as GC/MS or LC/MS/MS is required. Professional judgment should be applied to any drug test result, particularly when using preliminary positive results. For In Vitro Diagnostic Use Only.
The RapidFRET Oral Fluid PCP Calibrator Set and RapidFRET Oral Fluid PCP Control Set are intended for use only with the RapidFRET Oral Fluid Assay for PCP and samples collected with the RapidEASE Oral Fluid Collector. The cutoff calibrator is used to determine the cutoff level and translate the assay measurement into a positive or negative result. The positive and negative controls are used to monitor laboratory systems, operators, precision, accuracy and assay conditions. For In Vitro Diagnostic Use Only.
The RapidFRET Oral Fluid Assay for PCP is an In Vitro Diagnostic competitive immunoassay used to detect PCP in human oral fluid. This is a ready-to-use homogenous system that involves energy transfer between an acceptor fluorophore labeled to an antibody and a donor fluorophore labeled to drug. The assay is based on competition between drug in the sample and drug labeled with the donor fluorophore for a fixed number of binding sites on the antibody reagent. When acceptor and donor fluorophores are brought into close proximity through a binding event, energy transfer occurs. The fluorescence resonance energy transfer (FRET) signal is measured at the wavelength of the acceptor fluorophore and is inversely proportional to the amount of drug in the sample. A Cutoff Calibrator is used to translate the sample measurement into a positive or negative result. Controls are used to establish and monitor precision and accuracy. The assay is performed on the RapidFRET Integrated Workstation.
Here's a summary of the acceptance criteria and study details for the Biophor Diagnostics, Inc. RapidFRET Oral Fluid Assay for PCP, based on the provided text:
Acceptance Criteria and Device Performance
| Acceptance Criteria Category | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Precision | Demonstrates consistent and reproducible results across various PCP concentrations (0% to 200% of cutoff). | Successfully demonstrated across three lots over a minimum of 20 days. Results were consistently positive at and above 125% of cutoff and consistently negative at and below 100% of cutoff, with some variability around the 100% cutoff (2 positive, 277 negative). |
| Correlation with GC/MS (Accuracy) | High agreement with a confirmed analytical method (GC/MS). | Overall Agreement: >99% |
- RapidFRET POS / GC/MS POS: 119
- RapidFRET NEG / GC/MS NEG: 126
- RapidFRET POS / GC/MS NEG: 1 (Sample was 9 ng/mL PCP by GC/MS, which is below the 10 ng/mL assay cutoff)
- RapidFRET NEG / GC/MS POS: 0 |
| Cross-Reactivity | Minimal false positive results due to structurally related or unrelated compounds, OTC/prescription medications, and drugs of abuse. | Only 4-HydroxyPCP (620 ng/mL) and PCM (310 ng/mL) were found to cross-react below 10,000 ng/mL at concentrations equivalent to the cutoff in the absence of PCP, indicating high specificity. |
| Analytical Specificity (Interfering Substances) | No significant interference from common substances (foods, dental products, pH variations, biological components, medications). | All tested compounds and pH variations (HSA, ethanol, baking soda, whole blood, hemoglobin, hydrogen peroxide, sodium chloride, cholesterol, denture adhesive, ascorbic acid, bilirubin, IgA, IgG, IgM, mouthwash, cough syrup, cranberry juice, orange juice, toothpaste, chewing tobacco, cigarettes, chewing gum, hard candy, teeth whitening strips, cola, water, antacid, coffee, tea) at specified concentrations resulted in NEG when spiked with 5 ng/mL PCP and POS when spiked with 15 ng/mL PCP, indicating no significant interference. |
Study Details
-
Sample sizes used for the test set and the data provenance:
- Precision Study: The table shows totals of 279-294 samples per PCP concentration level for evaluating precision across different lots and days. This implies several hundred unique test measurements.
- Correlation with GC/MS: n = 246 neat oral fluid samples.
- Cross-Reactivity & Analytical Specificity: Not explicitly stated as one single test set size, but involves:
- 175 different compounds for cross-reactivity.
- A list of common substances and pH levels for analytical specificity.
- Data Provenance:
- "Neat oral fluid was collected with the RapidEASE Oral Fluid Collection Device from volunteers potentially positive and negative for PCP." This indicates prospective collection from volunteers.
- Country of origin is not explicitly stated, but the submission is to the U.S. FDA, typically implying data relevant to the U.S. market.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- Not applicable/mentioned for this type of in vitro diagnostic device study. The ground truth for drug concentration is established analytically, not through expert interpretation of images or clinical findings.
-
Adjudication method for the test set:
- Not applicable/mentioned for this type of in vitro diagnostic device study. Ground truth (PCP concentration) is determined by an objective, gold-standard chemical method (GC/MS or LC/MS/MS).
-
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, a multi-reader multi-case (MRMC) comparative effectiveness study was not performed. This device is an in vitro diagnostic assay, not an AI-assisted diagnostic tool that relies on human readers interpreting output. The read-out and interpretation (positive/negative) is automated based on the assay's cutoff.
-
If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
- Yes, the performance studies described are essentially standalone for the device. The RapidFRET Oral Fluid Assay for PCP, when run on the RapidFRET Integrated Workstation, provides an automated qualitative screen (positive/negative) for PCP. Human intervention is limited to sample collection, loading, and interpreting the instrument's P/N result (with subsequent confirmation by GC/MS or LC/MS/MS). The assay itself is an "algorithm only" in the sense that it mechanically and chemically determines the result based on a pre-defined cutoff.
-
The type of ground truth used:
- Analytical Confirmation (GC/MS): For the "Correlation with GC/MS" study, Gas Chromatography/Mass Spectrometry (GC/MS) was used as the confirmatory method to establish the true PCP concentration in the samples. The text also mentions LC/MS/MS as another specific alternate chemical method for confirmation. This represents a gold-standard analytical method.
-
The sample size for the training set:
- Not explicitly stated for a training set in the context of machine learning. This is an immunoassay, not a machine learning algorithm that requires a distinct training set. The development of the assay, its reagents, and its cutoff would be optimized through internal research and development, which implicitly involves testing various formulations and parameters to achieve desired performance characteristics.
-
How the ground truth for the training set was established:
- Not applicable in the context of a "training set" for a traditional immunoassay. The performance of the assay (e.g., sensitivity, specificity, cutoff) is established empirically through experimentation, using precisely prepared samples with known concentrations of PCP (spiked samples) and confirmed clinical samples (by GC/MS). The cutoff (10 ng/mL) is a predefined analytical threshold, not learned from a dataset.
N/A