Search Results

The xTAG® Gastrointestinal Pathogen Panel (GPP) is a multiplexed nucleic acid test intended for the simultaneous qualitative detection and identification of multiple viral, bacterial, and parasitic nucleic acids in human stool specimens or human stool in Cary-Blair media from individuals with signs and symptoms of infectious colitis or gastroenteritis. The following pathogen types, subtypes and toxin genes are identified using the xTAG GPP:

Viruses
. Adenovirus 40/41
Norovirus GI/GII ●
Rotavirus A

Bacteria
Campylobacter (C. jejuni, C. coli and C. lari only) ●
Clostridium difficile (C. difficile) toxin A/B
Escherichia coli (E. coli) 0157
Enterotoxigenic Escherichia coli (ETEC) LT/ST
. Salmonella
Shiga-like Toxin producing E. coli (STEC) stx 1/stx 2
Shigella (S. boydii, S. sonnei, S. flexneri and S. dysenteriae)
Vibrio cholerae (V. cholerae) cholera toxin gene (ctx) ●

Parasites
Cryptosporidium (C. parvum and C. hominis only) ●
Entamoeba histolytica (E. histolytica)
. Giardia (G. lamblia only, also known as G. intestinalis and G. duodenalis)

The detection and identification of 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, laboratory findings and epidemiological information. A gastrointestinal microorganism multiplex nucleic acid-based assay also aids in the detection and identification of acute gastroenteritis in the context of outbreaks.

xTAG GPP positive results are presumptive and must be confirmed by FDA-cleared tests or other acceptable reference methods.

The results of this test should not be used as the sole basis for diagnosis, treatment, or other patient management decisions. Confirmed positive results do not rule out co-infection with other organisms that are not detected by this test, and may not be the sole or definitive cause of patient illness. Negative xTAG GPP 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 noninfectious causes such as ulcerative colitis, irritable bowel syndrome, or Crohn's disease.

xTAG GPP is not intended to monitor or guide treatment for C. difficile infections.

The xTAG GPP test is indicated for use with the Luminex® 100/200™ and MAGPIX® instruments with xPONENT® software.

xTAG GPP incorporates a multiplex reverse-transcription polymerase chain reaction (RT-PCR) with Luminex's proprietary universal sorting system (the xTAG Universal Array) on the Luminex platform. The xTAG Universal Array sorts nucleic acids onto discreet Luminex bead populations by virtue of highly specific "tag/anti-tag" hybridization reactions. The tags and anti-tags comprising the xTAG Universal Array are 24-mer oligonucleotide sequences not found in nature. The assay has been designed to simultaneously detect microbial targets and an internal control (bacteriophage MS2 added to each sample prior to extraction).

For each sample, 10 µL of extracted nucleic acid is amplified in a single multiplex RT-PCR reaction. Amplimers ranging from 58 to 202 bp (not including the 24-mer tag) are generated in this reaction. A five μL aliquot of the RT-PCR product is then subjected to a hybridization/detection reaction that also includes bead populations coupled to 24-mer antitags. Each bead population is coupled to a unique anti-tag which is the exact complement of a 24-mer tag incorporated into a given amplimer. Thus, each Luminex bead population uniquely identifies a microbial target or assay control through a specific tag/anti-tag hybridization reaction. Signal is generated via a Streptavidin, R-Phycoerythrin conjugate.

The Luminex instrument sorts the products of these hybridization reactions and generates a signal in the form of a median fluorescence intensity (MFI) value for each bead population. The MFI values are generated by the xPONENT software provided with the instrument using the GPP protocol parameters, and are analyzed by the xTAG Data Analysis Software (TDAS GPP (US)). TDAS GPP (US) applies algorithms to MFI values in order to generate a qualitative result for each microbial target selected for reporting to establish the presence or absence of bacterial, viral or parasitic targets and/or controls in each sample. The data analysis software also generates a qualitative result and compiles a report for patient samples and external controls assayed in a given run. Before data are analyzed, a user has the option to select a subset of the targets from the intended use of the xTAG GPP (for each sample).

The provided text describes the Luminex xTAG® Gastrointestinal Pathogen Panel (GPP) and various studies supporting its performance. I will extract the requested information based on the additional analytes (Adenovirus 40/41, Entamoeba histolytica, and Vibrio cholerae) while noting that much of the original device's performance data (from K121894) is still applicable.

Description of Acceptance Criteria and Studies Proving Device Performance

The Luminex xTAG® Gastrointestinal Pathogen Panel (GPP) is a multiplexed nucleic acid test designed for the qualitative detection and identification of multiple viral, bacterial, and parasitic nucleic acids in human stool specimens or human stool in Cary-Blair media. This submission (K140647) primarily focuses on the addition of Adenovirus 40/41, Entamoeba histolytica, and Vibrio cholerae to the panel, as well as the use of stool in Cary-Blair media.

The device performance was evaluated through analytical studies (Analytical Reactivity, Carry-over Contamination, Limit of Detection, Repeatability, Analytical Specificity/Interference, Fresh vs. Frozen Stool Evaluation, Reproducibility/Precision) and clinical studies (Detection in Asymptomatic Volunteers, Detection in Symptomatic Patients from prospective and retrospective cohorts, and simulated samples).

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria for clinical performance are primarily inferred from the detailed reporting of Positive Percent Agreement (PPA) / Sensitivity and Negative Percent Agreement (NPA) / Specificity with their 95% Confidence Intervals (CI). Analytical performance criteria are noted qualitatively (e.g., "100% agreement," "no interference").

Clinical Performance Acceptance Criteria (Inferred from reporting, generally striving for high agreement):

• Sensitivity/Positive Agreement: Typically >90% with lower bound 95% CI >80% (explicitly stated for Norovirus GI/GII and Clostridium difficile toxin A/B in Cary-Blair media, and implied for other analytes by robust performance).
• Specificity/Negative Agreement: Typically >90% with lower bound 95% CI >90% (explicitly stated for all targets in Cary-Blair media, and implied for other analytes by robust performance).

2. Sample size used for the test set and the data provenance

• Asymptomatic Volunteers (Test Set): 200 clinical stool samples.
  • Data Provenance: Retrospective, collected from healthy, asymptomatic donors. Country of origin not specified, but collected for this study.
• Symptomatic Patients (Prospective Clinical Study in Stool Specimens) (Test Set): 1407 clinical stool specimens.
  • Data Provenance: Prospective, collected between June 2011 and February 2012. Origin: Four laboratories in the United States (Arizona, Missouri, Tennessee, Texas) and two in Southern Ontario (Canada).
• Symptomatic Patients (Prospective Clinical Study in Stool in Cary-Blair Media) (Test Set): A subset of the 1407 prospective samples for which adequate sample was available. The exact number varied per analyte for each comparison (e.g., Adenovirus 40/41 had N=6 positives in Cary-Blair for PPA calculation).
  • Data Provenance: Prospective, from the same cohort described above.
• Pre-selected stool specimens Retrospective Study (Test Set): 207 archived stool specimens positive for various pathogens (low prevalence in prospective study), plus negative clinical specimens.
  • Data Provenance: Retrospective, collected at multiple sites in North America, Africa, and Europe.
• Pre-selected Stool in Cary-Blair Specimens Retrospective Study (Test Set): Remnants of available pre-selected frozen stool specimens mixed proportionally with Cary-Blair medium.
  • Data Provenance: Retrospective, from the same pre-selected cohort.
• Botswana Pediatric Stool Specimens (Test Set): 313 pediatric stool specimens.
  • Data Provenance: Prospective, collected between February 2011 and January 2012 from symptomatic pediatric patients in Botswana, Africa.

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

The document does not explicitly state the number of experts or their specific qualifications (e.g., "radiologist with 10 years of experience") for establishing ground truth. Instead, it refers to "reference/comparator methods" or "comparator methods" performed at "central laboratories independent of xTAG GPP testing sites." For Entamoeba histolytica, microscopy was part of the comparator method, implying trained microscopists. For other analytes, molecular methods (EIA, amplification + sequencing, bacterial culture) were used, suggesting laboratory professionals proficient in these techniques.

4. Adjudication method for the test set

The document describes the establishment of ground truth for clinical studies using "Comparator methods."

• For Adenovirus 40/41: Composite comparator consisting of Premier Adenoclone Type 40/41 EIA and Amplification + sequencing.
• For Entamoeba histolytica: Microscopy followed by amplification + sequencing (for positive samples by microscopy only).
• For Vibrio cholerae: Bacterial culture.
• For other analytes (from K121894), similar molecular and culture-based comparator methods were used.

A "true positive" or "true negative" result for xTAG GPP was assigned if it agreed with the comparator method. In cases of discrepancies (e.g., xTAG GPP positive, comparator negative), confirmatory testing (e.g., bi-directional sequencing) was performed to determine if the xTAG GPP result was a true positive or false positive. For example, in the pre-selected study, confirmatory testing by NAAT and bi-directional sequencing was done on "additional positive calls" (xTAG GPP positive but not pre-selected).

This represents a form of discrepancy resolution, where a higher-tier reference method (like sequencing) is used to adjudicate disagreements. It is not an "expert consensus" adjudication in the sense of multiple independent experts reviewing and voting, but rather a technical adjudication against a more definitive laboratory standard.

5. 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 document describes a diagnostic device (multiplex nucleic acid test) for pathogen detection, not a system involving human readers interpreting images or data with and without AI assistance. Therefore, an MRMC comparative effectiveness study regarding human reader improvement with AI assistance is not applicable to this device and was not conducted. The "readers" here are automated instruments and software, and the "human-in-the-loop" refers to laboratory technicians performing the assay and interpreting the automated output based on predefined algorithms and cut-offs.

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

Yes, the analytical and clinical performance studies described represent a standalone (algorithm only) performance evaluation of the xTAG GPP device. The system automates nucleic acid extraction, amplification, detection, and analysis (TDAS GPP software applies algorithms to MFI values to generate qualitative results, which are then compiled into a report). The studies assess the accuracy of these automated results against comparator methods. While human operators perform sample preparation and run the instruments, the critical detection and identification steps, as well as the interpretation of raw data into qualitative positive/negative calls, are performed by the device's integrated system and software.

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

The ground truth used was a combination of:

• Reference Laboratory Methods / Comparator Methods: These included FDA-cleared EIAs, bacterial cultures, microscopy, and Nucleic Acid Amplification Tests (NAATs) followed by bi-directional sequencing for specific gene targets. These are often considered the gold standard or best available methods in microbiology diagnostics.
• Discrepancy Resolution using Sequencing: For discrepant results (e.g., xTAG GPP positive, comparator negative), bi-directional sequencing, which is a highly specific molecular method, was used as a definitive arbiter to confirm the presence or absence of the target nucleic acid.
• Clinical Outcomes Data: This was not explicitly stated as a direct method for establishing ground truth for individual pathogens within the scope discussed. The focus was on direct pathogen detection.

8. The sample size for the training set

The document does not explicitly delineate a "training set" in the context of machine learning. The studies described are for analytical validation and clinical performance evaluation of the completed device.

However, the "design history file" and "in silico analysis" mentioned (e.g., for primer cross-reactivity) imply that analytical methods were developed and refined based on bioinformatics and laboratory testing during the device's development phase. The analytical reactivity studies, for instance, involve empirical testing of a wide range of strains to establish reactivity, which could be seen as part of the developmental data used to "train" or optimize the assay's ability to detect targets. The document focuses on the validation data for the regulatory submission.

9. How the ground truth for the training set was established

As noted above, a formal "training set" with established ground truth, as understood in machine learning, is not described. The analytical studies (e.g., Limit of Detection, Analytical Reactivity) used well-characterized strains, genotypes, isolates, or clinical specimens with known positivity/negativity established by commercial sources (e.g., ATCC, Zeptometrix) or validated laboratory methods (e.g., quantitative PCR, cell culture infectious dose determination). This type of well-characterized material would be analogous to ground truth for developing and optimizing the assay's performance characteristics.

Ask a specific question about this device

The xTAG® Gastrointestinal Pathogen Panel (GPP) is a multiplexed nucleic acid test intended for the simultaneous qualitative detection and identification of multiple viral, bacterial, and parasitic nucleic acids in human stool specimens or human stool in Cary-Blair media from individuals with signs and symptoms of infectious colitis or gastroenteritis. The following pathogen types, subtypes and toxin genes are identified using the xTAG GPP:

Viruses

• . Adenovirus 40/41
• Norovirus GI/GII ●
• Rotavirus A

Bacteria

• Campylobacter (C. jejuni, C. coli and C. lari only) .
• Clostridium difficile (C. difficile) toxin A/B ●
• Escherichia coli (E. coli) 0157
• Enterotoxigenic Escherichia coli (ETEC) LT/ST
• . Salmonella
• Shiga-like Toxin producing E. coli (STEC) stx 1/stx 2 ●
• Shigella (S. boydii, S. sonnei, S. flexneri and S. dysenteriae)
• Vibrio cholerae (V. cholerae) cholera toxin gene (ctx)

Parasites

• Cryptosporidium (C. parvum and C. hominis only) .
• Entamoeba histolytica (E. histolytica) ●
• . Giardia (G. lamblia only - also known as G. intestinalis and G. duodenalis)

The detection and identification of 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, laboratory findings and epidemiological information. A gastrointestinal microorganism multiplex nucleic acid-based assay also aids in the detection and identification of acute gastroenteritis in the context of outbreaks.

xTAG GPP positive results are presumptive and must be confirmed by FDA-cleared tests or other acceptable reference methods.

The results of this test should not be used as the sole basis for diagnosis, treatment, or other patient management decisions. Confirmed positive results do not rule out co-infection with other organisms that are not detected by this test, and may not be the sole or definitive cause of patient illness. Negative xTAG Gastrointestinal Pathogen Panel 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.

xTAG GPP is not intended to monitor or guide treatment for C. difficile infections.

The xTAG GPP is indicated for use with the Luminex® 100/200™ instrument system with xPONENT® software.

xTAG GPP incorporates a multiplex reverse-transcription polymerase chain reaction (RT-PCR) with Luminex's proprietary universal sorting system (the xTAG Universal Array) on the Luminex platform. The xTAG Universal Array sorts nucleic acids onto discreet Luminex bead populations by virtue of highly specific "tag/anti-tag" hybridization reactions. The tags and anti-tags comprising the xTAG Universal Array are 24-mer oligonucleotide sequences not found in nature. The assay has been designed to simultaneously detect microbial targets and an internal control (bacteriophage MS2 added to each sample prior to extraction).

For each sample, 10 µL of extracted nucleic acid is amplified in a single multiplex RT-PCR reaction. Amplimers ranging from 58 to 202 bp (not including the 24-mer tag) are generated in this reaction. A five μL aliquot of the RT-PCR product is then subjected to a hybridization/detection reaction that also includes bead populations coupled to 24-mer antitags. Each bead population is coupled to a unique anti-tag which is the exact complement of a 24-mer tag incorporated into a given amplimer. Thus, each Luminex bead population uniquely identifies a microbial target or assay control through a specific tag/anti-tag hybridization reaction. Signal is generated via a Streptavidin, R-Phycoerythrin conjugate.

The Luminex instrument sorts the products of these hybridization reactions and generates a signal in the form of a median fluorescence intensity (MFI) value for each bead population. The MFI values are generated by the xPONENT software provided with the instrument using the GPP protocol parameters, and are analyzed by the xTAG Data Analysis Software (TDAS GPP (US)). TDAS GPP (US) applies algorithms to MFI values in order to generate a qualitative result for each microbial target selected for reporting to establish the presence or absence of bacterial, viral or parasitic targets and/or controls in each sample. The data analysis software also generates a qualitative result and compiles a report for patient samples and external controls assayed in a given run. Before data are analyzed, a user has the option to select a subset of the targets from the intended use of the xTAG GPP (for each sample).

Here's a summary of the acceptance criteria and study information for the xTAG Gastrointestinal Pathogen Panel (GPP) device, based on the provided text:

Acceptance Criteria and Device Performance

The acceptance criteria are generally established as ranges or thresholds for performance metrics. For specific analytes, detailed ranges for Mean MFI and %CV are provided as device performance, particularly in the reproducibility studies for Low Positive (LP) and Moderate Positive (MP) samples. Overall, the qualitative agreement with expected results and confidence intervals (95% CI) are the primary indicators of acceptance for both analytical and clinical performance.

Note on Table Content: Due to the extensive nature of the data, a comprehensive single table with all acceptance criteria and reported performance for every analyte across all studies would be extremely large. The table below focuses on general acceptance criteria mentioned for clinical performance and provides examples of reported performance for the additional analytes emphasized in this submission (Adenovirus 40/41, Entamoeba histolytica, Vibrio cholerae). For full details on all analytes from the original submission (K121454), one would need to consult that document.

Study Details

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

   • Analytical Reactivity: Empirical testing of a wide range of clinically relevant GI pathogen strains, genotypes, and isolates. For new analytes, specific strains (e.g., Adenovirus 40/41 from Zeptometrix, CDC; E. histolytica from ATCC; V. cholerae from NCTC, ATCC) were used. Additionally, sequencing analysis was performed on 9 Adenovirus 40 and 28 Adenovirus 41 clinical samples.
   • Carry-over Contamination: 144 High Negative samples and 144 High Positive samples for Adenovirus 40, run in a checkerboard manner in duplicate after 6 independent extractions.
   • Limit of Detection (LoD): Serial dilutions of simulated samples in negative clinical stool matrix, confirmed with at least 20 replicates of the selected dilution for each analyte.
   • Repeatability: 20 replicates for each of two different analyte concentrations (LoD and 5x-10x above cut-off MFI).
   • Analytical Specificity: Organisms (cross-reactivity, interference) tested at high positive titers, panel analytes at low positive concentrations. Specific numbers of organisms are not consistently provided for all tests but mentioned for E. dispar strains (ATCC PRA-353, PRA-368).
   • Fresh vs. Frozen Stability: Sets of simulated specimens prepared at Low Positive, Moderate Positive, and High Positive concentrations. Specific numbers are not explicitly stated for each level but indicate replicates were run.
   • Precision/Reproducibility: 90 replicates for each single analyte and dual analyte sample (3 replicates per run x 5 runs per operators per site x 3 sites).
   • Stool in Cary-Blair Media LoD Study: Serial dilutions for Clostridium difficile, Giardia lamblia, Norovirus GII (representative analytes). 20 replicates for each LoD titer.
   • Asymptomatic Volunteers (Clinical Study): 200 clinical stool samples. Data was analyzed for 192 samples due to PCR inhibition.
   • Prospective Clinical Study (Stool): 1407 clinical specimens. Collected prospectively between June 2011 and February 2012 from pediatric and adult patients. Data provenance: 4 laboratories in the United States (Arizona, Missouri, Tennessee, Texas) and 2 in Southern Ontario (Canada).
   • Prospective Clinical Study (Stool in Cary-Blair): A subset of samples from the 1407 prospective study for which adequate sample was available.
   • Pre-selected Stool Specimens (Retrospective Study): 207 archived stool specimens. Collected retrospectively from multiple sites in North America, Africa, and Europe.
   • Pre-selected Stool in Cary-Blair Specimens (Retrospective Study): Remnants of available pre-selected frozen stool specimens mixed proportionally with Cary-Blair.
   • Supplemental Clinical Data (Simulated Stool/Cary-Blair): 50 negative, 50 positive E. histolytica, 50 positive V. cholerae simulated stool specimens. 50 negative, 50 positive Adenovirus 40/41, 50 positive E. histolytica, 50 positive V. cholerae simulated stool in Cary-Blair specimens.
   • Botswana Pediatric Stool Specimens (Retrospective Study): 313 pediatric stool specimens. Collected prospectively between February 2011 and January 2012 from symptomatic pediatric patients in Botswana, Africa.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience):

   The document does not explicitly state the number or specific qualifications (e.g., years of experience) of experts involved in establishing ground truth. However, it indicates:

   • Reference/Comparator methods: For the clinical studies, reference/comparator methods were used at central laboratories independent of xTAG GPP testing sites. These methods included Premier Adenoclone Type 40/41 EIA, Amplification + sequencing (one NAAT+), Microscopy, and Bacterial culture.
   • Sequencing analysis: Bi-directional sequencing with analytically validated primers was performed. This implies expertise in molecular biology and sequence analysis.
   • Clinical evaluation: The study was conducted at 6 independent laboratories (4 in US, 2 in Canada) for the prospective study, and 4 sites for the retrospective study (3 external to Luminex Molecular Diagnostics - LMD). This suggests that the ground truth assessment, particularly for clinical diagnosis and interpretation of comparator methods, involved multiple trained laboratory personnel and clinicians.

3. Adjudication method (e.g. 2+1, 3+1, none) for the test set:

   The document describes confirmatory testing for discrepant results, particularly in the retrospective and Botswana studies. This often implies an adjudication process:

   • For positive xTAG GPP results not pre-selected at the banking site, confirmatory testing by nucleic acid amplification followed by bi-directional sequencing was performed to determine if these were True Positive or False Positive.
   • For certain analytes in the Botswana study, a random subset of xTAG GPP negative samples was also assessed by nucleic acid amplification and bi-directional sequencing.
   • One specimen that was positive for Adenovirus 40/41 by comparator but negative by xTAG GPP was positive by bi-directional sequencing only (i.e. FDA-cleared EIA negative).
   • These actions suggest a form of discrepant analysis, where a more definitive "gold standard" (like sequencing) is used to adjudicate differences. However, a formal "m+n" type adjudication (e.g., 2+1, 3+1) involving a panel of experts for every case is not explicitly stated. The comparator methods themselves serve as the primary reference standard, with sequencing as a confirmatory tool for discrepancies.

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 device is a diagnostic assay (in vitro diagnostic device), not an AI-assisted imaging or diagnostic tool that involves human readers interpreting results with or without AI assistance. Therefore, a multi-reader multi-case (MRMC) comparative effectiveness study focusing on human reader improvement with AI vs. without AI assistance was not performed. The evaluation is focused on the standalone performance of the assay itself compared to existing reference methods.

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

   Yes, a standalone performance evaluation was done. The entire analytical and clinical performance studies described (analytical reactivity, LoD, repeatability, specificity, stability, reproducibility, clinical prospective, retrospective, and supplemental studies) are all assessments of the algorithm-only performance of the xTAG GPP device. The device automates the detection and identification of nucleic acids from pathogens, and the results are qualitatively interpreted by the xTAG Data Analysis Software (TDAS GPP (US)) which applies algorithms to MFI values to generate a "positive" or "negative" result. Human involvement is in sample preparation, running the instrument, and reviewing the generated report, but the "performance" discussed is that of the assay and its associated software for detection, not human interpretation aided by AI.

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

   The ground truth was primarily established through reference/comparator methods, which varied by pathogen:

   • Composite Comparator: For Adenovirus 40/41 (Premier Adenoclone Type 40/41 EIA + Amplification + sequencing).
   • Bacterial Culture: For Vibrio cholerae.
   • Microscopy followed by Amplification + Sequencing: For Entamoeba histolytica (positive specimens by microscopy only).
   • Nucleic Acid Amplification Tests (NAATs) followed by bi-directional sequencing: Used for confirmatory testing of discrepant results across various analytes.
   • In-house real-time PCR: Used for some confirmations (e.g., Giardia, Adenovirus).
     These methods represent established diagnostic assays and molecular techniques, often considered "gold standards" or highly accurate reference methods in clinical microbiology.

7. The sample size for the training set:

   The document describes various analytical and clinical studies, but it does not explicitly mention a "training set" in the context of machine learning model development. This is typical for traditional in vitro diagnostic devices. The performance evaluation focuses on validation, not iterative training and testing of an AI model.

   However, if we broadly interpret "training set" as data used during the development and initial optimization before formal validation:

   • The document states that the reagents were unchanged from a previous submission (K121454). This implies development and some level of internal testing and optimization (akin to a training phase) would have occurred prior to that submission using various pathogen strains and clinical samples to establish the assay's parameters.
   • The "Analytical Reactivity" studies involve testing a wide range of strains to establish reactivity, which could be considered part of an iterative development process that feeds into refinement.

   Without specific details on Luminex's internal development process, a distinct "training set" with a defined sample size for an AI algorithm is not identified in this regulatory submission.

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

   Since a discrete "training set" for an AI algorithm is not explicitly identified, the method for establishing ground truth for such a set is also not detailed.
   However, for the general development and optimization of the assay:

   • Known strains and characterized clinical specimens: Analytical studies like analytical reactivity and LoD determination use well-characterized, often commercially available, strains (e.g., ATCC, Zeptometrix, NCTC) and high-titer clinical specimens where the presence and concentration of the pathogen are already known through established microbiological and molecular techniques. These serve as the "ground truth" for calibrating and optimizing assay parameters (like MFI cut-offs) during development.
   • Reference methods: Any prior development would have relied on the same types of reference methods (culture, EIA, PCR/sequencing) that were used to establish ground truth in the formal validation studies.

Ask a specific question about this device

The xTAG® Gastrointestinal Pathogen Panel (GPP) is a multiplexed nucleic acid test intended for the simultaneous qualitative detection and identification of multiple viral, parasitic, and bacterial nucleic acids in human stool specimens from individuals with signs and symptoms of infectious colitis or gastroenteritis. The following pathogen types, subtypes and toxin genes are identified using the xTAG GPP:

• Campylobacter (C. jejuni, C. coli and C. lari only) .
• Clostridium difficile (C. difficile) toxin A/B ●
• Cryptosporidium (C. parvum and C. hominis only) ●
• Escherichia coli (E. coli) 0157 .
• Enterotoxigenic Escherichia coli (ETEC) LT/ST ●
• Giardia (G. lamblia only also known as G. intestinalis and G. duodenalis)
• . Norovirus GI/GII
• . Rotavirus A
• Salmonella ●
• Shiga-like Toxin producing E. coli (STEC) stx 1/stx 2 .
• Shigella (S. boydii, S. sonnei, S. flexneri and S. dysenteriae) ●

The detection and identification of 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, laboratory findings and epidemiological information. A gastrointestinal microorganism multiplex nucleic acid-based assay also aids in the detection and identification of acute gastroenteritis in the context of outbreaks.

xTAG® GPP positive results are presumptive and must be confirmed by FDAcleared tests or other acceptable reference methods.

The results of this test should not be used as the sole basis for diagnosis, treatment, or other patient management decisions. Confirmed positive results do out rule out coinfection with other organisms that are not detected by this test, and may not be the sole or definitive cause of patient illness. Negative xTAG® Gastrointestinal Pathogen Panel 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.

xTAG GPP is not intended to monitor or guide treatment for C. difficile infections.

The xTAG GPP is indicated for use with the Luminex® MAGPIX® instrument.

The xTAG® Gastrointestinal Pathogen Panel (GPP) is a multiplexed nucleic acid test.

Here's an analysis of the provided text, outlining the acceptance criteria and study details for the xTAG® Gastrointestinal Pathogen Panel (GPP):

Device Name: xTAG® Gastrointestinal Pathogen Panel (GPP)
Regulation Number: 21 CFR 866.3990
Regulation Name: Gastrointestinal microorganism multiplex nucleic acid-based assay
Regulatory Class: Class II
Product Code: PCH, NSU, JJH

This document is a marketing clearance (510(k)) letter from the FDA, granting substantial equivalence to the xTAG® GPP. While it defines the device and its intended use, it does not contain the detailed study results, acceptance criteria, or performance data typically found in the submission package. The letter refers to the submission document (K121894) where this information would be located.

Therefore, many of the requested details cannot be extracted directly from the provided text. I will indicate where the information is not available in this document.

Acceptance Criteria and Study to Prove Device Meets Acceptance Criteria

1. Table of Acceptance Criteria and Reported Device Performance

• Acceptance Criteria: Not explicitly stated in this document.
• Reported Device Performance: Not explicitly stated in this document.

To fulfill this section, one would typically need access to the original 510(k) submission (K121894) or associated peer-reviewed publications. The performance data would usually involve sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for each pathogen detected by the panel, compared against a reference method.

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size for Test Set: Not available in this document.
• Data Provenance (e.g., country of origin, retrospective/prospective): Not available in this document.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

• Number of Experts: Not available in this document.
• Qualifications of Experts: Not available in this document.

4. Adjudication Method for the Test Set

• Adjudication Method: Not available in this document.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was Done

• MRMC Study: This type of study (comparing human readers with and without AI assistance) is typically relevant for interpretative diagnostic devices, especially those involving image analysis. The xTAG® GPP is a multiplex nucleic acid test, which is an in vitro diagnostic (IVD) device that directly detects pathogens from a sample. Therefore, an MRMC study is not applicable to this type of device. The test itself generates a result (positive/negative for specific pathogens), not an image for human interpretation that would be "assisted" by AI.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done

• Standalone Performance: For an IVD like the xTAG® GPP, the "algorithm" is the entire assay process (sample preparation, nucleic acid extraction, amplification, detection using xTAG technology, and data analysis software). The performance is inherently standalone in terms of generating a result from a sample. The device performance itself (sensitivity, specificity) reported in the 510(k) would represent this standalone performance. However, the specific metrics are not available in this document.

7. The Type of Ground Truth Used

• Type of Ground Truth: While not explicitly stated here, for an IVD diagnostic device, the ground truth for performance studies is typically established by FDA-cleared reference methods (e.g., culture, PCR, or other gold-standard diagnostic tests) or a composite reference standard (combining multiple methods, sometimes with expert review) for each pathogen. The document states: "xTAG® GPP positive results are presumptive and must be confirmed by FDA-cleared tests or other acceptable reference methods." This implies that the ground truth used in validation studies would have relied on such reference methods.

8. The Sample Size for the Training Set

• Sample Size for Training Set: Not applicable in the traditional sense for this type of IVD. This device is a molecular diagnostic assay, not a machine learning model that undergoes "training" on a dataset in the same way an image recognition AI would. The "training" of such a device involves assay development, optimization, and analytical validation. However, there would be validation sample sizes, which are not available in this document.

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

• Ground Truth for Training Set: As above, the concept of a "training set" and its "ground truth" establishment in the context of an IVD is different from AI/ML. Assay development and analytical validation would use samples characterized by established reference methods to ensure the assay performs as expected. The details are not available in this document.

Summary of Missing Information from the Provided Document:

The provided FDA clearance letter (K121894) confirms the device's regulatory status and indications for use but does not contain the detailed technical performance data, study design, or ground truth establishment methods. This information would be found in the original premarket notification submission itself.

Ask a specific question about this device