Search Results

The BD MAX™ Cdiff Assay performed on the BD MAX™ System is an automated in vitro diagnostic test for the direct, qualitative detection of the Clostridium difficile toxin B gene (tcdB) in human liquid or soft stool specimens from patients suspected of having C. difficile infection (CDI). The test. performed directly on the specimen, utilizes real-time polymerase chain reaction (PCR) for the amplification of C. difficile toxin B gene DNA and fluorogenic target-specific hybridization probes for the detection of the amplified DNA. The BD MAX ™ Cdiff Assay is intended to aid in the diagnosis of CDI.

The BD MAX™ System and the BD MAX™ Cdiff Assay are comprised of an instrument with associated hardware and accessories, disposable microfluidic cartridges, master mixes, unitized reagent strips, extraction reagents, and sample buffer tubes. The instrument automates sample preparation including target Ivsis. DNA extraction and concentration, reagent rehydration, and target nucleic acid amplification and detection using real-time PCR. The assay includes a Sample Processing Control (SPC) that is present in the Extraction Tube. The SPC monitors DNA extraction steps, thermal cycling steps, reagent integrity and the presence of inhibitory substances. The BD MAX™ System software automatically interprets test result may be called as NEG (negative), POS (positive) or UNR (unresolved) based on the amplification status of the target and of the Sample Processing Control. IND (indeterminate) or INC (incomplete) results are due to BD MAX™ System failure.

BD MAX™ Cdiff Assay - Acceptance Criteria and Study Findings

The BD MAX™ Cdiff Assay is an automated in vitro diagnostic test for the direct, qualitative detection of the Clostridium difficile toxin B gene (tcdB) in human liquid or soft stool specimens. The assay is intended to aid in the diagnosis of C. difficile infection (CDI).

1. Table of Acceptance Criteria and Reported Device Performance

The provided document details various analytical and clinical performance studies, but explicit "acceptance criteria" for minimum performance thresholds (e.g., "Sensitivity must be >= 90%") are not stated in a direct, consolidated table. However, the reported performance metrics can be inferred as meeting the implicit acceptance criteria for the submission to be cleared.

Here's a summary of the reported device performance from the provided text, focusing on key metrics:

2. Sample Size and Data Provenance

• Test Set (Clinical Performance Studies):

  • Clinical Performance (vs. Reference Method): 1819 compliant and reportable results from 2071 initial soft or liquid stool specimens.
  • Clinical Performance (vs. Direct Culture): 1826 compliant and reportable results.
  • Comparison Studies: 2013 specimens tested across two external sites (specific breakdown per comparator assay: 647 for one and 1366 for the other).
  • Data Provenance: Multi-site prospective investigational study from six investigational centers. The document does not specify countries of origin, but the submitting company is Canadian. The study was prospective, as specimens were from "patients suspected of having C. difficile infection for which diagnostic tests were indicated and ordered."

• Training Set: The document does not provide specific details on a separate "training set" for algorithm development. This is common for molecular diagnostic assays where the "training" (e.g., selection of primers/probes, optimization of PCR conditions) is inherent in the assay's design and analytical validation, rather than a distinct machine learning training phase.

3. Number of Experts and Qualifications for Ground Truth

• Ground Truth for Clinical Performance (Reference Method): The ground truth was established by a "Comparative Reference Method consisting of direct culture complemented by enriched culture." The anaerobic culture was used to isolate C. difficile, followed by confirmation of isolate identification and a "Tissue Culture Cytotoxicity Assay to determine the toxigenicity of the isolate." This process typically involves trained microbiology laboratory personnel, but the specific number and qualifications of "experts" (e.g., clinical microbiologists, laboratory directors) involved in establishing the final ground truth from this multi-step process are not explicitly stated in the document.

• Ground Truth for Analytical Studies (e.g., LoD, Inclusivity, Specificity): For analytical studies, ground truth was based on known concentrations of bacterial strains (CFU/mL or genomic copies/mL) or known presence/absence of certain organisms, established through standard laboratory methods of quantification and identification. No "experts" in the sense of clinical reviewers are typically involved in establishing these analytical truths.

4. Adjudication Method for the Test Set

The document does not describe a formal adjudication method (e.g., 2+1, 3+1 reviewer consensus) by human experts for the clinical test set results. The ground truth for the clinical performance was established by a "Comparative Reference Method" (direct culture + enriched culture + toxigenicity testing). Discordant results between the BD MAX™ Cdiff Assay and this Reference Method were further investigated:

• "27 of 48 False Positive BD MAX™ Cdiff specimens were also found to be positive using another commercially available FDA-cleared RT-PCR assay targeting the C. difficile tcdB gene."
• "32 of 37 False Negative BD MAX™ Cdiff specimens, were also found to be negative using another commercially available FDA-cleared RT-PCR assay targeting the C. difficile tcdB gene."
  This "further investigation" serves as a form of resolution for discordant cases but is not a traditional multi-reader "adjudication" process using human interpretation of images or clinical data.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No Multi-Reader Multi-Case (MRMC) comparative effectiveness study was described. The device is an automated in vitro diagnostic test that provides a qualitative (positive/negative) result, meaning there is no human "reader" in the loop for interpreting the assay's output that would then be compared to human performance with or without AI assistance. The study compares the device's performance against a reference method and other FDA-cleared molecular assays.

6. Standalone Performance

Yes, a standalone (algorithm only without human-in-the-loop performance) study was done. The entire clinical and analytical performance evaluation describes the performance of the BD MAX™ Cdiff Assay as a standalone device. The assay automates sample preparation, nucleic acid amplification, and detection, with the BD MAX™ System software automatically interpreting test results as NEG (negative), POS (positive), or UNR (unresolved).

7. Type of Ground Truth Used

• Clinical Performance Studies: The ground truth was based on a "Comparative Reference Method" which primarily relied on culture-based methods (direct and enriched anaerobic culture to isolate C. difficile) combined with a Tissue Culture Cytotoxicity Assay to determine the toxigenicity of the isolated strains. This represents a combination of microbiological culture and a functional assay (outcome of bacterial toxin activity on cells).
• Analytical Performance Studies (LoD, Inclusivity, Specificity): The ground truth was based on known characterized bacterial strains and their concentrations, determined by standard microbiological quantification methods.

8. Sample Size for the Training Set

The document does not specify a distinct "training set" sample size in the context of the device's development or a machine learning algorithm. For molecular diagnostic assays like this, the "training" involves the scientific development and optimization of the assay components (primers, probes, reaction conditions) performed by molecular biologists and R&D scientists, rather than a separate data-driven training phase for an AI model.

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

As described above, no explicit "training set" with established ground truth in the machine learning sense is detailed. The assay's design and optimization (analogous to "training" in AI development) would have utilized known positive and negative controls, characterized bacterial strains, and various concentrations to perfect the real-time PCR parameters and ensure accurate detection of the tcdB gene. This process involves standard laboratory validation techniques rather than human expert annotation for ground truth.

Ask a specific question about this device