(169 days)
The BD GeneOhm™ Cdiff Assay is a rapid in vitro diagnostic test for the direct, qualitative detection of C. difficile toxin B gene (todB) in human liquid or soft stool specimens from patients suspected of having Clostridium difficile-associated disease (CDAD). The test, based on real-time PCR, is intended for use as an aid in diagnosis of CDAD. The test is performed directly on the specimen, utilizing polymerase chain reaction (PCR) for the amplification of specific targets and fluorogenic target-specific hybridization probes for the detection of the amplified DNA.
The BD GeneOhm™ Cdiff Assay is a rapid in vitro diagnostic test for the direct, qualitative detection of C. difficile toxin B gene (todB) in human liquid or soft stool specimens from patients suspected of having Clostridium difficile-associated disease (CDAD). The test, based on real-time PCR, is intended for use as an aid in diagnosis of CDAD. The test is performed directly on the specimen, utilizing polymerase chain reaction (PCR) for the amplification of specific targets and fluorogenic target-specific hybridization probes for the detection of the amplified DNA.
A liquid or soft stool specimen is collected and transported to the laboratory. A sterile dry swab is dipped into the liquid or soft stool material and processed. For testing, the swab is eluted in sample buffer and the specimen is lysed. An aliquot of the lysate is added to PCR readents which contain the tcdB specific primers used to amplify the genetic target of Clostridium difficile, if present. The assay also includes an internal control (IC) to detect PCR inhibited specimens and to confirm the integrity of assay reagents. Amplified targets are detected with hybridization probes labelled with quenched fluorophores (molecular beacons). The amplification, detection and interpretation of the signals are done automatically by the Cepheid SmartCycler® software. The entire procedure takes about 75 to 90 minutes, depending on the number of specimens processed.
Here's a breakdown of the acceptance criteria and the study details for the BD GeneOhm™ Cdiff Assay, based on the provided 510(k) summary:
Acceptance Criteria and Device Performance
The acceptance criteria are implied by the reported performance metrics. The device needed to demonstrate sufficient sensitivity, specificity, and reproducibility compared to a recognized reference method to be considered substantially equivalent.
| Performance Metric | Acceptance Criteria (Implied) | Reported Device Performance (Overall) - Fresh Samples | Reported Device Performance (Overall) - Frozen Samples |
|---|---|---|---|
| Sensitivity | High (e.g., >90%) | 93.8% (95% CI: 86.2% - 98.0%) | 100.0% (95% CI: 89.7% - 100%) |
| Specificity | High (e.g., >95%) | 95.5% (95% CI: 93.8% - 96.9%) | 97.7% (95% CI: 94.8% - 99.3%) |
| Negative Predictive Value (NPV) | High (e.g., >95%) | 99.1% | 99.2% |
| Positive Predictive Value (PPV) | Reasonably High | 67.3% | 81.5% |
| Unresolved Rate (Overall) | Low (e.g., 95%) | Site-to-Site: 96.7%, Lot-to-Lot: 100.0% | Not explicitly broken down for frozen/fresh |
| Reproducibility (Moderate Positive Agreement) | High (e.g., >95%) | Site-to-Site: 100%, Lot-to-Lot: 97.8% | Not explicitly broken down for frozen/fresh |
| Reproducibility (Negative Agreement) | High (e.g., >95%) | Site-to-Site: 100%, Lot-to-Lot: 100.0% | Not explicitly broken down for frozen/fresh |
| Interfering Substances | No detectable interference | Most tested substances showed no interference (NI = No Interference) | Not applicable (qualitative) |
Details of the Study
- Sample Size and Data Provenance:
- Clinical Study (Test Set):
- Total Specimens Tested: 1108
- Reportable Results: 1090
- Fresh Specimens (First Dataset): 835 specimens
- Frozen Specimens (Second Dataset): 255 specimens (retested from aliquots of original stool specimens after an issue with the reference assay at one site)
- Data Provenance: Multi-site prospective investigational study conducted at four medical centers (two in Canada and two in the United States).
- Clinical Study (Test Set):
- Number of Experts and Qualifications for Ground Truth (Clinical Study):
- The document does not specify the number of experts used or their qualifications for establishing the ground truth directly. It refers to the "Reference Cytotoxicity Assay" as the comparator method.
- Adjudication Method (Clinical Study):
- The document does not explicitly describe an adjudication method for discrepancies between the device and the reference assay in the clinical study. It notes that for 21 out of 34 samples that were positive by the reference assay and negative by the BD GeneOhm™ Cdiff Assay, further testing (Cytotoxicity Assay on isolated strains, standard PCR with alternative primers, and bi-directional sequencing) was performed to verify the presence of toxigenic C. difficile. This implies a form of ad-hoc adjudication for discordant results rather than a pre-defined process for all cases.
- Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
- No, an MRMC comparative effectiveness study involving human readers with and without AI assistance was not performed. This device is an in vitro diagnostic (IVD) assay, not an AI-assisted diagnostic tool for human interpretation. The "interpretation of the signals are done automatically by the Cepheid SmartCycler® software."
- Standalone Performance:
- Yes, the performance characteristics reported (sensitivity, specificity, NPV, PPV) are for the device (BD GeneOhm™ Cdiff Assay) in standalone mode, directly compared against the reference method. There is no human-in-the-loop performance described for the primary clinical effectiveness assessment.
- Type of Ground Truth Used (Clinical Study):
- The primary ground truth for the clinical study was the Reference Cytotoxicity Assay on liquid or soft stool specimens. For discordant results, further molecular testing (standard PCR with alternative primers and bi-directional sequencing) was used to verify the presence of the tcdB gene or toxigenic C. difficile. This indicates a combination of phenotypic (cytotoxicity) and genotypic (molecular) methods to establish the most accurate ground truth.
- Sample Size for the Training Set:
- The document describes performance characteristics of a completed assay rather than an algorithm trained on a dataset. Therefore, there is no explicit "training set" sample size mentioned in the context of machine learning model development. The analytical studies (specificity, sensitivity, reproducibility) served to characterize the assay's performance against known samples.
- How Ground Truth for the Training Set Was Established:
- As there's no explicitly defined "training set" in the context of a machine learning algorithm, this question is not directly applicable. For the analytical studies:
- Analytical Specificity: Strains were identified and confirmed as non-target organisms or C. difficile lacking the tcdB gene. The "neg" result indicates the absence of the target gene.
- Analytical Sensitivity (LOD): Used a specific C. difficile strain (ATCC 43255) with known concentrations (DNA copies/CFU) and confirmed with other toxigenic strains. This involved careful preparation and quantification of known positive samples.
- Reproducibility: Involved simulated specimens, with "low positive," "moderate positive," and "negative" categories established by inoculating known quantities of C. difficile (ATCC 43255) into simulated bowel flora.
- As there's no explicitly defined "training set" in the context of a machine learning algorithm, this question is not directly applicable. For the analytical studies:
§ 866.2660 Microorganism differentiation and identification device.
(a)
Identification. A microorganism differentiation and identification device is a device intended for medical purposes that consists of one or more components, such as differential culture media, biochemical reagents, and paper discs or paper strips impregnated with test reagents, that are usually contained in individual compartments and used to differentiate and identify selected microorganisms. The device aids in the diagnosis of disease.(b)
Classification. Class I (general controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to § 866.9.