LogoFDA 510(k) Search
K Number
K111928
Device Name
PSYCHEMEDICS PHENCYCLIDINE EIA
Manufacturer
PSYCHEMEDICS CORP.
Date Cleared
2012-05-01

(300 days)

Product Code
LCM
Regulation Number
N/A
Type
Traditional
Panel
CH
Reference & Predicate Devices
k011275
AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
Intended Use

The Psychemedics Microplate EIA for Phencyclidine is an enzyme immunoassay (EIA) for the preliminary qualitative detection of phencyclidine in human head and body hair samples using a phencyclidine calibrator at 3 ng /10 mg hair cutoff for the purpose of identifying phencyclidine use. This is an in vitro diagnostic device intended exclusively for Psychemedics use only and is not intended for sale to anyone. The Psychemedics EIA Phencyclidine Assay provides only a preliminary analytical test result. To obtain a quantitative analytical result or to confirm positive results, a more specific alternate chemical method (e.g. GC/MS) must be used. Clinical consideration and professional judgment should be applied to the interpretation of any drug-of-abuse test result.

Device Description

The test consists of two parts; a pre-analytical hair treatment procedure (to convert the solid matrix of hair to a measurable liquid matrix) and the screening assay, the Psychemedics Microplate EIA for Phencyclidine. The drug is recovered from the hair using a patented method (U.S. Patent #8.084.215).The screening portion of the test system consists of (1) microplate wells coated with multiple drugs including phencyclidine conjugated to bovine serum albumin (BSA) (patent pending), polyclonal rabbit anti-phencyclidine, goat anti-rabbit secondary antibody conjugated to HRP (horseradish peroxidase), substrate [3, 3', 5, 5' tetramethylbenzidine (TMB)], HCl to acidify the final reaction, and wash buffer for washing the plates. Absorbance in the wells is read with a microplate reader.

AI/ML Overview

Here's a breakdown of the acceptance criteria and the study details for the Psychemedics Microplate EIA for Phencyclidine in Hair, based on the provided 510(k) summary:

Acceptance Criteria and Reported Device Performance

Acceptance Criteria CategorySpecific Acceptance Criteria (Implied/Direct)Reported Device Performance
Agreement with GC/MSHigh agreement (e.g., sensitivity, specificity, accuracy) for preliminary qualitative detection of phencyclidine at the 3 ng/10 mg hair cutoff.Concordance with GC/MS:
  • EIA Positive, GC/MS ≤ -10% of Cutoff: 0
  • EIA Positive, GC/MS Between -10% and -50% of Cutoff: 0
  • EIA Positive, GC/MS Between -50% and Cutoff: 2
  • EIA Positive, GC/MS Between Cutoff and +50%: 8
  • EIA Positive, GC/MS Between +50% and +100%: 7
  • EIA Positive, GC/MS > +100% of Cutoff: 46
  • EIA Negative, GC/MS ≤ -10% of Cutoff: 140
  • EIA Negative, GC/MS Between -10% and -50% of Cutoff: 1
  • EIA Negative, GC/MS Between -50% and Cutoff: 13
  • EIA Negative, GC/MS Between Cutoff and +50%: 0
  • EIA Negative, GC/MS Between +50% and +100%: 0
  • EIA Negative, GC/MS > +100% of Cutoff: 0 |
    | Agreement with Predicate Device | High agreement in classifying samples as positive or negative compared to the predicate device (Psychemedics Phencyclidine Assay, K011275). | Agreement with Predicate:
  • EIA Positive, Predicate Negative: (Data missing in table)
  • EIA Positive, Predicate Positive: (Data missing in table, likely refers to 'C' in table)
  • EIA Negative, Predicate Negative: 364
  • EIA Negative, Predicate Positive: (Data missing in table) |
    | Precision | Consistent results (intra-assay and inter-assay) across different concentrations relative to the cutoff (negative control, various percentages above/below cutoff). | Intra-Assay:
  • B₀ (-100%): 15 NEG, 0 POS
  • -75%: 15 NEG, 0 POS
  • -50%: 15 NEG, 0 POS
  • -25%: 15 NEG, 0 POS
  • +25%: 0 NEG, 15 POS
  • +50%: 0 NEG, 15 POS
  • +75%: 0 NEG, 15 POS
  • +100%: 0 NEG, 15 POS
    Inter-Assay:
  • B₀ (-100%): 75 NEG, 0 POS
  • -75%: 75 NEG, 0 POS
  • -50%: 75 NEG, 0 POS
  • -25%: 75 NEG, 0 POS
  • +25%: 0 NEG, 75 POS
  • +50%: 0 NEG, 75 POS
  • +75%: 0 NEG, 75 POS
  • +100%: 0 NEG, 75 POS |
    | Cosmetic Treatment Effects | Cosmetic treatments (bleach, permanent wave, dye, relaxer, shampoo) should not convert PCP-negative samples to positive, nor convert PCP-positive samples to negative. | Negative Samples: All 20 samples per treatment (bleach, perm, dye, relaxer, shampoo) remained negative after treatment (total 100 samples).
    Positive Samples: No samples positive for PCP became negative after cosmetic treatment (12 samples per treatment type, total 60 samples). Average B/B₀ x 100 values before and after treatment showed minimal change. |
    | Contamination Effects | The device should accurately identify contamination and distinguish it from actual drug use (ingestion). | PCP in Water Contamination:
  • 8 samples soaked in 1000 ng phencyclidine/mL water.
  • Post-wash, remaining PCP ranged from 3.3 to 16.4 ng/10 mg hair.
  • After applying the wash criterion, all were determined contaminated, not positive for ingestion.
    PCP in Saline Contamination:
  • 8 samples soaked in 1000 ng phencyclidine/mL saline.
  • Post-wash, remaining PCP ranged from 0.7 to 3.0 ng/10 mg hair.
  • Seven samples were negative without wash criterion.
  • One sample at cutoff was determined contaminated after applying wash criterion. |
    | Cross-reactivity | Minimal cross-reactivity with structurally similar compounds, and no cross-reactivity with a wide range of other substances. | Cross-reactivity observed:
  • 1-(1-Phenylcyclohexyl)morpholine (PCM): 5.0% (at 60 ng/10 mg hair)
  • Metaphit: 30% (at 10 ng/10 mg hair)
    No cross-reactivity: 64 other compounds.
    No interference with 128 compounds at +/- 50% of cutoff. |
    | Recovery | Recovery of PCP from hair of PCP users should be substantially equivalent to the predicate device. | Reported as "substantially equivalent to the method of the predicate device." |
    | Calibrator/Control Stability | Calibrator and control solutions should remain stable for a specified period. | Stability of PCP in methanol in the presence of other drugs of abuse shown to exceed 1 year. |

Study Details

  1. Sample sizes used for the test set and the data provenance:

    • Agreement Testing (compared to GC/MS):
      • Total samples: 229 samples (140 negative, 24 below cutoff, 15 between cutoff and +100% of cutoff, 46 >100% of cutoff).
      • Body hair samples: 15% of the total samples.
    • Agreement Testing (compared to Predicate Device):
      • Total samples: An additional 224 negative samples were compared to the predicate. (There appear to be missing values/errors in the provided table for positive comparisons).
    • Precision Studies:
      • Intra-Assay: 15 replicates per concentration level (total 8 levels tested).
      • Inter-Assay: 75 replicates per concentration level (total 8 levels tested).
    • Cosmetic Treatment Studies:
      • Negative samples: 100 samples (20 per treatment type: bleach, permanent wave, dye, relaxer, shampoo; 10 samples per brand for each treatment).
      • Positive samples: 60 samples (12 per treatment type: bleach, permanent wave, dye, relaxer, shampoo; 6 samples per brand for each treatment).
    • Contamination Study: 16 hair samples (8 soaked in water-PCP, 8 soaked in saline-PCP).
    • Cross-reactivity and Interference Studies: 66 compounds for cross-reactivity (2 showed cross-reactivity, 64 did not), and 128 compounds for interference.
    • Data Provenance: Not explicitly stated, but the submission is from a US company (Psychemedics Corporation, Culver City, CA). The context of a 510(k) submission generally implies prospective collection for performance evaluation or retrospective analysis of samples typical for the intended use.
    • Retrospective/Prospective: Not explicitly stated. The nature of the studies (e.g., controlled spiking for precision, controlled cosmetic treatments, contamination studies) suggests a mix of controlled experimental prospective studies and likely retrospective collection of clinical samples for agreement testing.

  2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

    • The ground truth for the test set (primarily for agreement studies) was established by GC/MS (Gas Chromatography/Mass Spectrometry), which is an analytical chemical method, not human expert consensus. GC/MS is considered the gold standard for confirming drug presence and concentration in forensic and clinical toxicology. Therefore, no human experts were used to establish this analytical ground truth.

  3. Adjudication method for the test set:

    • Not applicable/mentioned for the analytical ground truth (GC/MS). For the comparison between the new device and the predicate, there's no mention of an adjudication process; it's a direct comparison of results.

  4. 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 is an in vitro diagnostic (IVD) device, specifically an immunoassay for drug detection. It does not involve human readers interpreting results in the way an imaging AI typically would. Therefore, an MRMC comparative effectiveness study involving human readers and AI assistance is not applicable to this device.

  5. If a standalone (i.e. algorithm only without human-in-the loop performance) was done:

    • Yes, the performance studies described (Agreement with GC/MS, Precision, Cosmetic Treatment Effects, Contamination Effects, Cross-reactivity) represent the standalone performance of the Psychemedics Microplate EIA for Phencyclidine. The device provides a preliminary analytical test result. While a human is involved in running the assay and interpreting the preliminary qualitative result, the performance data presented are inherent to the assay's chemical and enzymatic reactions (the "algorithm" in a broad sense for IVDs). However, it's crucial to note the stated intended use that positive results must be confirmed by GC/MS, implying a human-in-the-loop for definitive diagnosis, but the reported performance metrics are for the assay itself.

  6. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

    • Analytical Ground Truth: GC/MS (Gas Chromatography/Mass Spectrometry) was used as the ground truth for confirming the presence and concentration of phencyclidine in hair samples. This is a highly accurate chemical method, considered the definitive standard for such measurements.

  7. The sample size for the training set:

    • The document describes device performance studies for a submitted medical device (Psychemedics Microplate EIA). These types of submissions typically focus on validation studies (test sets) rather than detailing the full development and "training" of the assay, especially since it's a traditional immunoassay, not a machine learning algorithm in the modern sense. Therefore, no specific "training set" sample size is mentioned or applicable in the context of this traditional immunoassay device. The assay design and formulation would have been optimized during development, but this wouldn't be referred to as a "training set" in the way it is for AI.

  8. How the ground truth for the training set was established:

    • As stated above, the concept of a "training set" and associated "ground truth" doesn't directly apply here as it would for an AI/ML-based device. The assay components and their interactions (e.g., antibodies, substrates, reagents) are designed based on known biochemical principles and then validated through performance testing against established analytical methods like GC/MS and predicate devices.

N/A