(266 days)
The BioCode Gastrointestinal Pathogen Panel (GPP) is a qualitative, gastrointestinal microorganism multiplexed nucleic acid-based assay capable of detecting of nucleic acids from the following organisms in unpreserved stool and Cary-Blair media:
- Adenovirus 40/41 ●
- Campylobacter (C. jejuni, C. coli) ●
- Clostridium difficile (C. difficile) toxin A/B (from fresh specimens only) ●
- Cryptosporidium (C. parvum, C. hominis) 0
- Entamoeba histolytica ●
- Escherichia coli (E. coli) 0157 ●
- Enterotoxigenic E. coli (ETEC) LT/ST ●
- Enteroaggregative E. coli (EAEC) 0
- Giardia lamblia (also known as G. intestinalis and G. duodenalis) ●
- Norovirus GI/GII ●
- Rotavirus A ●
- Salmonella spp. .
- Shiga-like Toxin producing E. coli (STEC) stx1/stx2 ●
- Shigella (S. boydii, S. sonnei, S. flexneri, S. dysenteriae)/EIEC ●
- Vibrio spp. (V. cholerae, V. parahaemolyticus, V. vulnificus), specific identification of V. ● parahaemolyticus
- Yersinia enterocolitica
The BioCode Gastrointestinal Pathogen Panel (GPP) is indicated as an aid in the diagnosis of specific agents of gastrointestinal illness and results are meant to be used in conjunction with other clinical, laboratory, and epidemiological data. Positive results do not rule out co-infection with organisms not included in the BioCode Gastrointestinal Pathogen Panel (GPP). The agent detected may not be the cause of patient illness. Negative results in the setting of clinical illness compatible with gastroenteritis may be due to infection by pathogens that are not detected by this test or non-infectious causes such as ulcerative colitis, irritable bowel syndrome, or Crohn's disease.
Concomitant culture is necessary for organism recovery and further typing of bacterial agents. This device is not intended to monitor or guide treatment for C. difficile infection.
Due to the small number of positive specimens collected for certain organisms during the prospective clinical study, performance characteristics for Campylobacter spp., E. coli 0157, Shigella/EIEC, Yersinia enterocolitica, and Adenovirus 40/41 were established primarily with retrospective clinical specimens.
Performance characteristics for Entamoeba histolytica, and Vibrio spp. (V. parahaemolyticus, V. vulnificus, and Vibrio cholerae) were established primarily using contrived clinical specimens.
The BioCode Gastrointestinal Pathogen Panel (GPP) is a multiplex nucleic acid test designed to be used with the BioCode MDx 300 system. The BioCode MDx 3000 is an automated system that integrates PCR amplification, target capture, signal generation and optical detection for nultiple gastrointestinal pathogens from a single stool speciment or in Cary Blair. Stool specimens are processed and nucleic acids extracted with easyMAG, an automated system. Once the PCR plate is set up and sealed, all other operations are automated on MDx 3000. The BioCode Gastrointestinal Pathogen Panel simultaneously tests for 17 pathogens (see table below) from unpreserved stool specimens or stool collected in Cary-Blair transport medium. Results from the BioCode Gastrointestinal Pathogen Panel test are available within less than 5 hours.
Here's a breakdown of the acceptance criteria and study details for the BioCode Gastrointestinal Pathogen Panel (GPP) based on the provided document:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state pre-defined acceptance criteria for the clinical performance metrics (PPA and NPA). Instead, it presents the calculated agreement rates and their 95% confidence intervals from the clinical study. The reproducibility study, however, does have clear acceptance criteria defined (>= 95% detection for positive targets, and >= 95% non-detection for negative targets).
Here's a table summarizing the reported clinical performance. Note that "Acceptance Criteria" for clinical performance are inferred based on general expectations for diagnostic assays submitted to the FDA; explicit numerical targets were not provided in the text for these. The Reproducibility section does state explicit criteria.
| Category | Target (General) | Acceptance Criteria (Explicit if available, otherwise inferred) | Reported Device Performance |
|---|---|---|---|
| Clinical Performance | Positive Percent Agreement (PPA) | Generally high agreement (e.g., >80-90% for most targets) | See tables on pages 14-17 for specific organism PPAs and 95% CIs. |
| Negative Percent Agreement (NPA) | Generally high agreement (e.g., >95% for most targets) | See tables on pages 14-17 for specific organism NPAs and 95% CIs. | |
| Reproducibility Study | Detection of positive targets (1.5x LoD and 3x LoD) | >= 95% detection | All specified targets at both concentrations achieved 100% detection, except STEC (98.89% at Low). All >95%. (pg 32) |
| Non-detection of negative targets (for negative control) | >= 95% non-detection | Giardia lamblia: 99.78% non-detection. All others 100%. All >95%. (pg 32) | |
| Limit of Detection | Lowest concentration with >=95% detection for 20 replicates | >= 95% detection (19/20) at specified LoD (for 20 replicates) | All targets met 20/20 or 19/20 detection at their determined LoD for 20 replicates across both unpreserved and Cary-Blair stool. (pg 35-36) |
| Fresh vs. Frozen Stability | 95% replicates for positive targets should be valid and detected (>=60/63) | 63/63 replicates for positive targets were valid and detected. (pg 50) | |
| 95% of negative targets should be valid and not detected (>=60/63) | 63/63 replicates for negative targets were valid and not detected (one was detected but invalid for RNA IC). (pg 50) | ||
| Interfering Substances/Microbes | No inhibition or unexpected results | No inhibition or unexpected results observed, with some exceptions/details noted for specific cross-reactivity findings (Vibrio spp., Yersinia enterocolitica, Shigella dysenteriae, Rotavirus vaccines, C. meleagridis). (pg 37-38) | |
| Analytical Reactivity/Inclusivity | Detection of diverse strains at 3X LoD | Detection at 3X LoD for most strains; higher if not. | All organisms were detected at the indicated concentrations. (pg 38-45) |
| Analytical Specificity/Cross-Reactivity | No cross-reactivity with common commensals/pathogens | No cross-reactivity observed/predicted with list of organisms, with some exceptions/details noted. (pg 46-48) | |
| Competitive Inhibition | No competitive inhibition observed | No competitive inhibition observed. (pg 48) | |
| Cross Contamination/Sample Carryover | Absence of carryover or cross-contamination | No cross-contamination was observed. (pg 50) |
2. Sample Sizes and Data Provenance
-
Clinical Study (Prospective):
- Sample Size: 1558 leftover, de-identified samples.
- Data Provenance: Prospectively collected from patients in the United States (Baltimore, MD; Tampa, FL; Memphis, TN; Los Angeles, CA). Samples collected between January 2015 and August 2017.
- Specimen Breakdown:
- Unpreserved (Fresh): 237 samples
- Unpreserved (Frozen): 960 samples
- Cary-Blair (Fresh): 361 samples
- Inoculated Cary-Blair Samples: 400 unpreserved stool samples from Sites 1 and 2 were thawed and inoculated into Cary-Blair to supplement numbers for specific analytes (pg 21).
-
Pre-selected Archived Specimens (Category III):
- Sample Size: 260 preselected archived specimens.
- Data Provenance: Archived clinical specimens that were previously tested positive by different methods (retrospective). These were spiked with known positives and randomized with negative specimens. (pg 23)
-
Contrived Specimens (Category IV):
- Sample Size: 612 samples total, with 485 positive samples.
- Data Provenance: Prepared using specimens previously tested negative for all BioCode GPP analytes, then spiked at levels of up to 3X LOD or greater. (pg 25)
-
Asymptomatic Volunteers (Clinical Specificity):
- Sample Size: 125 clinical stool samples.
- Data Provenance: Collected from healthy asymptomatic donors from Tampa General Hospital and University of Maryland. (pg 30-31)
-
Reproducibility Study:
- Sample Size: 7 contrived samples (6 positive, 1 negative control) extracted in triplicate and each assayed in singlet over 10 runs per site. Total 90 replicates per target. (pg 32)
- Data Provenance: Performed at 3 sites. (pg 32)
-
Limit of Detection (LoD):
- Sample Size: 20 replicates for each tested organism and sample type (unpreserved stool and Cary-Blair). (pg 35)
-
Interfering Substances:
- Each sample extracted in triplicate and tested in singlet. (pg 37)
-
Cross Contamination/Sample Carryover:
- Five complete runs in "checkerboard" pattern. (pg 50)
3. Number of Experts and Qualifications for Ground Truth
The document does not specify the number or qualifications of experts used to establish ground truth for the clinical test set. Instead, it relies on reference methods for clinical performance evaluation (pg 13).
4. Adjudication Method for the Test Set
The document does not describe an adjudication method involving human experts. For the clinical study, the reference method for each target pathogen was used as the ground truth. In some cases, discrepancies between the BioCode GPP and the initial reference method were subjected to bidirectional sequencing or additional rounds of sequencing for "confirmation" (e.g., footnotes on pages 14, 15, 17, 22), which can be considered a form of analytical adjudication to determine the true state of the specimen.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No MRMC study was mentioned. The device is a "multiplex nucleic acid-based assay," indicating an automated laboratory test, not an imaging AI diagnostic that assists human readers. Therefore, the concept of "human readers improve with AI vs without AI assistance" does not apply here.
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance)
Yes, the studies presented are for the standalone performance of the BioCode Gastrointestinal Pathogen Panel (GPP) system, without human interpretation or assistance beyond typical laboratory operational procedures. The results, as quantitative fluorescence intensity (MFI) values, are processed by the BioCode MDx 3000 system to determine positive or negative results.
7. Type of Ground Truth Used
-
Clinical Study (Prospective):
- Reference standard methods: Culture, FDA-cleared NAAT (Nucleic Acid Amplification Test), or Composite results of two PCR/sequencing assays (pg 13).
- For discrepancies, bidirectional sequencing or additional rounds of sequencing were used as further confirmation/adjudication (e.g., footnotes on pages 14, 15, 17, 22).
-
Archived Specimens: Previously tested positive by different methods, and verified using analyte-specific PCR followed by bi-directional sequencing performed by Applied BioCode, Inc. (pg 23).
-
Contrived Specimens: Known concentrations of quantified bacteria, viruses, or parasites spiked into negative clinical matrix. (pg 25)
8. Sample Size for the Training Set
The document does not explicitly mention a "training set" in the context of machine learning. This device is a molecular diagnostic assay, not a machine learning algorithm that is trained on a dataset. The studies described are performance validation studies for an established assay.
9. How the Ground Truth for the Training Set Was Established
Since this is not a machine learning device trained on a "training set," this question is not applicable. The device's probes and primers for detecting targets are designed based on known genetic sequences, not through a data-driven training process like that used for AI/ML.
§ 866.3990 Gastrointestinal microorganism multiplex nucleic acid-based assay.
(a)
Identification. A gastrointestinal microorganism multiplex nucleic acid-based assay is a qualitativein vitro diagnostic device intended to simultaneously detect and identify multiple gastrointestinal microbial nucleic acids extracted from human stool specimens. The device detects specific nucleic acid sequences for organism identification as well as for determining the presence of toxin genes. The detection and identification of a specific gastrointestinal microbial nucleic acid from individuals exhibiting signs and symptoms of gastrointestinal infection aids in the diagnosis of gastrointestinal infection when used in conjunction with clinical evaluation and other laboratory findings. A gastrointestinal microorganism multiplex nucleic acid-based assay also aids in the detection and identification of acute gastroenteritis in the context of outbreaks.(b)
Classification. Class II (special controls). The special controls are set forth in FDA's guideline document entitled: “Class II Special Controls Guideline: Gastrointestinal Microorganism Multiplex Nucleic Acid-Based Assays for Detection and Identification of Microorganisms and Toxin Genes from Human Stool Specimens.” For availability of the guideline document, see § 866.1(e).