Search Results

The LifeScale AST system is a multiplexed in vitro diagnostic test that uses a microfluidic sensor and resonant frequency to calculate organism concentration and/or mass distribution for quantitative antimicrobial susceptibility testing (AST). Testing is performed directly on blood cultures signaled as positive by a continuous monitoring blood culture system and confirmed by Gram stain. The LifeScale AST system does not provide organism identification and is not indicated for use with polymicrobial samples. Interpretive results (Susceptible/Intermediate/Susceptible-dose dependent/Resistant) are provided for specific drug/organism combinations. Results are intended to be used in conjunction with other clinical and laboratory findings. Standard laboratory protocols for processing positive blood cultures should be followed to ensure availability of isolates for supplemental testing as needed. Additionally, subculture of positive blood culture is necessary for the susceptibility testing of organisms present in polymicrobial samples, for testing antimicrobial agents and species not indicated for testing with the device and for epidemiologic testing and for recovery of organisms present in microbial samples.

The LifeScale Gram Negative Kit (LSGN) is intended for use with the LifeScale AST system for in vitro testing of positive blood culture samples confirmed by Gram stain as containing gram-negative bacilli for the antimicrobial agents and specific target organisms identified below:

1. Ampicillin: Escherichia coli
2. Aztreonam: Escherichia coli, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca
3. Cefazolin: Escherichia coli, Klebsiella pneumoniae, Klebsiella variicola
4. Ceftazidime: Acinetobacter spp. (other than Acinetobacter ursingii), Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola, Klebsiella pneumoniae, Pseudomonas aeruginosa
5. Ertapenem: Escherichia coli, Klebsiella aerogenes, Klebsiella pneumoniae, Klebsiella oxytoca
6. Trimethoprim-Sulfamethoxazole: Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola
7. Amikacin: Acinetobacter spp., Escherichia coli, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola, Pseudomonas aeruqinosa
8. Cefepime: Escherichia coli, Klebsiella pneumoniae, Klebsiella oxytoca, Pseudomonas aeruginosa
9. Ceftazidime-avibactam: Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca
10. Gentamicin: Escherichia coli, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola, Pseudomonas aeruginosa
11. Levofloxacin: Escherichia coli, Klebsiella pneumoniae, Klebsiella derogenes, Klebsiella oxytoca, Pseudomonas aeruginosa
12. Meropenem: Acinetobacter spp., Escherichia coli, Klebsiella pneumoniae, Klebsiella oxytoca, Pseudomonas aeruginosa
13. Meropenem-vaborbactam: Escherichia coli, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca
14. Piperacillin-tazobactam: Acinetobacter spp., Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa

The Affinity Biosensors LifeScale Gram Negative Kit (LSGN) is a semi-automated instrument system for antimicrobial susceptibility testing (AST) directly from positive blood cultures for which the Gram stain shows gram-negative bacilli. The system uses a microfluidic sensor that detects organisms in suspension and measures differences in cell mass between bacterial suspensions incubated in the presence and absence of antibiotic. Minimum inhibitory concentrations (MICs) are determined from data obtained during sample measurement including organism concentration and/or cell mass distributions of individual organisms. The system automatically interprets the measurements to determine MIC values and interpretive results (susceptible, intermediate, or resistant) based on FDA-defined or recognized breakpoints. The organism identification determined using a platform FDA-cleared for use with positive blood culture samples is entered by the user. If the organism identification has not been entered or if the sample has not been confirmed as monomicrobial, the system provides a preliminary report that indicates that organism identification or monomicrobial status is pending. The device Instructions for Use indicates that the preliminary laboratory report should not be reported to the healthcare provider. The final report is provided to the healthcare provider when the organism identification is entered into the system and the culture is confirmed to be monomicrobial samples should not be tested with the LifeScale LSGN Kit. Preliminary results are available in most cases within four hours from initiation of the assay.

Here's a summary of the acceptance criteria and the study details for the LifeScale Gram Negative Kit (LSGN) with the LifeScale AST system, based on the provided FDA 510(k) submission:

1. Table of Acceptance Criteria and Reported Device Performance (Clinical Performance Data)

The acceptance criteria for clinical performance are implicitly demonstrated through the achieved Essential Agreement (EA), Category Agreement (CA), Very Major Discrepancy (VMJ), Major Discrepancy (MAJ), and Minor Discrepancy (MIN) rates. While specific pass/fail thresholds for each of these are not explicitly stated as "acceptance criteria" in a single table, the summary of performance data (Table 3) and subsequent detailed descriptions for each antimicrobial agent confirm the performance was considered acceptable, with a few specified limitations. Generally, for most AST devices, high EA and CA (typically >90-95%) and low VMJ/MAJ rates (typically 89%) reproducibility" was stated for the reproducibility study itself.

For inoculum density, the study aimed to show that organism concentration does not impact performance and that the device terminates tests with insufficient growth.

For media equivalence, an EA agreement of "≥95% for the various blood culture media compared to LifeScale AST mode MICs from the BD BACTEC Standard Aerobic Media" was the acceptance criterion.

For interfering substances, "EA ≥ 95%" was acceptable.

Below is a summarized table of clinical performance for the additional claimed antimicrobial/organism combinations and the performance for the removal of limitations, as these were the focus of the submission. The table combines various performance metrics from Tables 3 and 12, focusing on the overall performance for each drug/organism group. Specific criteria are often tied to regulatory guidelines (e.g., CLSI, FDA), which this document notes were used.

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

The test set consisted of both prospective clinical blood cultures (PBCs) and contrived blood cultures.

• Clinical Samples: 986 total clinical tests were initiated (Table 2). The specific breakdown by drug/organism combination for clinical samples is embedded within the detailed performance sections (e.g., "465 clinical (75.5%)" for Amikacin, indicating 465 clinical samples for Amikacin).
  • Provenance: Collected at 6 US Clinical sites.
  • Retrospective/Prospective: Prospective.
• Contrived Samples: A total of 3307 analytical tests were initiated, which includes contrived samples (Table 2). For individual drug/organism combinations, the number of challenge samples (e.g., "151 challenge (24.5%)" for Amikacin) contributes to the overall number of samples evaluated.
  • Provenance: Prepared from frozen isolates supplied by Affinity Biosensors or contemporary isolates collected by the laboratory.
  • Retrospective/Prospective: Primarily prospective (prepared and tested during the study).

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

The document does not explicitly state the number of experts or their qualifications for establishing ground truth. However, it indicates:

• Ground Truth Method: The reference method for Antimicrobial Susceptibility Testing (AST) was Broth Microdilution (BMD), performed according to CLSI guidance (CLSI M07).
• Organism Identification: Organism ID for prospective samples was performed using an FDA-cleared direct from positive blood culture ID system, and confirmed by MALDI (matrix-assisted laser desorption/ionization). If there was a discordant identification, MALDI was considered the final identification.
• Reference Testing Location: Reference testing was performed at two trial sites that received isolates from the PBC purity panel.
• This suggests that trained laboratory personnel with expertise in clinical microbiology and adherence to CLSI standards established ground truth.

4. Adjudication Method for the Test Set

The document states:

• "Each sample submitted for BMD testing was assigned a unique Trial ID, and LifeScale results were kept blinded to prevent bias."
• Performance was evaluated by comparing quantitative (MIC) and qualitative (S/I/SDD/R) AST results generated by the LifeScale AST System with those of the reference BMD.
• No explicit "adjudication method" (like 2+1 or 3+1 expert review) for resolving discrepancies between the device and ground truth is described. The comparison appears to be a direct assessment against the BMD reference. Discrepancies (VMJ, MAJ, MIN) are reported, but a formal adjudication process for these discrepancies by additional experts is not detailed.

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

No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done. The study focuses on the performance of the automated device (LifeScale AST system) compared to a laboratory reference method (BMD), rather than comparing human reader performance with and without AI assistance. This device is an automated AST system, not an AI-assisted diagnostic tool for human readers.

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

Yes, the study primarily assessed standalone performance of the LifeScale AST system. The system automatically interprets measurements to determine MIC values and interpretive results. While a user enters the organism identification, the core AST determination is an automated function of the device, making this a standalone performance assessment in the context of AST. The phrase "algorithm only without human-in-the loop performance" applies to the AST result generation itself.

7. The Type of Ground Truth Used

The primary ground truth used was FDA-recognized reference Broth Microdilution (BMD) method, following CLSI guidance (CLSI M07). For organism identification, MALDI was the definitive method.

8. The Sample Size for the Training Set

The document does not provide information on the sample size for the training set. This submission is a 510(k) for an addition of claims to an already cleared device (K211815). The focus is on the performance evaluation for the new claims and removal of limitations, implying that the underlying algorithms were likely developed and validated prior to this specific submission.

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

The document does not provide information on how the ground truth for the training set was established, as it pertains to the development of the device's algorithms. As mentioned above, this document focuses on the performance of the device for new claims, not its initial development.

Ask a specific question about this device

The VITEK® REVEAL™ AST System is an automated system for qualitative antimicrobial susceptibility testing (AST) of organisms direct from positive blood culture. The VITEK® REVEAL™ AST System does not provide organism identification.

The VITEK® REVEAL™ AST System is an automated system that uses an array of sensors to detect volatile organic compounds emitted by growing bacteria for the in vitro quantitative determination of antimicrobial susceptibility. The VITEK® REVEAL™ GN AST Assay is indicated for susceptibility testing direct from positive blood culture samples signaled as positive by a continuous monitoring blood culture system and confirmed to contain gramnegative bacilli by Gram stain. Organism identification is required for AST result interpretation and reporting.

This test is performed by laboratory health professionals in a clinical diagnostic setting. Results may be used as an aid to clinicians in determining appropriate antimicrobial therapy. Test results from the VITEK® REVEAL™ AST System should be interpreted in conjunction with other clinical and laboratory findings. Standard laboratory protocols for processing positive blood cultures should be followed to ensure availability of isolates for supplemental testing. Sub-culturing is necessary to support further testing for: bacteria and antimicrobials not on the VITEK® REVEAL™ GN AST Assay panel, inconclusive results, epidemiologic testing, recovery of organisms present in positive blood cultures samples, and susceptibility testing of bacteria in polymicrobial samples.

The VITEK® REVEAL™ GN AST Assay tests the following antimicrobial agents with the specific target organisms identified below:

Amikacin: Acinetobacter baumannii-calcoaceticus complex, Citrobacter freundii (including Citrobacter freundii complex), Enterobacter cloacae complex, Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group, Proteus mirabilis, Pseudomonas aeruginosa, Serratia marcescens

Amoxicillin/clavulanate: Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group, Proteus mirabilis

Ampicillin/sulbactam: Escherichia coli, Klebsiella oxytoca, Proteus mirabilis

Aztreonam: Citrobacter freundii complex), Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae (including K. pneumoniae group), Pseudomonas aeruginosa

Cefepime: Citrobacter koseri (syn. C. diversus), Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella species (including K. pneumoniae group and K. aerogenes), Klebsiella oxytoca, Pseudomonas aeruginosa

Cefotaxime: Acinetobacter baumanni-calcoaceticus complex, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae group

Ceftazidime: Acinetobacter baumannii-calcoaceticus complex, Citrobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae group

Ceftazidime/avibactam: Citrobacter freundii complex, Citrobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella aerogenes, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis, Pseudomonas aeruginosa

Ceftolozane/tazobactam: Citrobacter koseri, Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa

Ceftriaxone: Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis

Cefuroxime: Citrobacter koseri, Escherichia coli, Klebsiella pneumoniae group, Klebsiella oxytoca, Proteus mirabilis

Ciprofloxacin: Citrobacter freundit complex), Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens

Ertapenem: Escherichia coli, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis, Proteus vulgaris

Gentamicin: Citrobacter freundii complex, Citrobacter koseri, Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens

Imipenem: Acinetobacter baumannii-calcoaceticus complex, Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group, Pseudomonas aeruginosa, Serratia marcescens

Levofloxacin: Citrobacter koseri, Citrobacter freundii complex), Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens

Meropenem: Acinetobacter baumannii-calcoaceticus complex, Enterobacter cloacae (including E. cloacae complex), Escherichia coli, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens

Meropenem/vaborbactam: Citrobacter freundii (incluidng C. freundii complex), Citrobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae (including K. pneumoniae group), Proteus mirabilis

Piperacillin/tazobactam: Citrobacter koseri, Escherichia coli, Klebsiella pneumoniae group), Proteus vulgaris

Tetracycline: Acinetobacter baumannii-calcoaceticus complex, Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group

Tobramycin: Citrobacter freundii complex, Citrobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae group, Proteus mirabilis, Pseudomonas aeruginosa, Serratia marcescens

Trimethoprim/sulfamethoxazole: Escherichia coli, Klebsiella aerogenes, Klebsiella pneumoniae group

ESBL Confirmation test: Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae group

The VITEK® REVEAL™ AST System is an in vitro diagnostic (IVD) automated platform for phenotypic Antimicrobial Susceptibility Testing (AST) of bacterial samples, directly from positive blood cultures. The System utilizes broth microdilution (BMD) principles to quickly and accurately determine Minimum Inhibitory Concentrations (MIC) for the drugs on the VITEK® REVEAL™ GN AST Assay, and in combination with species identification (obtained from an FDA-cleared rapid ID method), will provide a Susceptible / Intermediate / Resistant (SIR) determination, or a Positive/Negative (POS/NEG) determination for the ESBL Confirmation screen test, for the species tested. The VITEK® REVEAL™ AST System is indicated for susceptibility testing of specific Gramnegative bacteria commonly associated with bacteremia (Table 1).

Sample preparation for testing in the VITEK® REVEAL™ AST System is fast, simple, and requires minimal skill. After a blood culture sample is identified as positive by a validated, automated blood culture system, a Gram stain is performed to confirm positivity and to determine whether the sample is Gram-positive, Gram-negative, or yeast. Samples determined by Gram stain to be monomicrobial for Gram-negative bacteria are diluted in Pluronic water and dispensed into VITEK® REVEAL™ Antibiotic Panels, containing serial dilutions of antibiotics and dried media. A VITEK® REVEAL™ Sensor Panel is sealed atop an inoculated VITEK® REVEAL™ Antibiotic Panel using the VITEK® REVEAL™ Sealer in an AST disposable assembly comprising a VITEK® REVEAL™ GN AST Assay.

The VITEK® REVEAL™ AST System detects bacterial growth using an array of proprietary chemical Small Molecule Sensors (SMS), which change color in the presence of various metabolic gases (volatile organic compounds) emitted by growing bacteria during incubation. The SMS arrays, printed onto the VITEK® REVEAL™ Sensor Panel, are positioned atop each well of the VITEK® REVEAL™ Antibiotic Panel. The sealed VITEK® REVEAL™ GN AST Assay is placed in the VITEK® REVEAL™ Instrument, which functions as an incubator for the samples being tested and optically monitors and tracks the change in sensor colors as the bacteria grow. These color changes are monitored by a scan every 10 minutes, allowing a real-time assessment of growth as a function of antibiotic concentration. A real-time algorithm detects sensor array responses indicating the volatile-compound emissions that are associated with bacterial population growth. Each antimicrobial agent-containing well is then compared to the response in control wells (the positive control well containing no antimicrobial agent, and the negative containing no growth media). Bacterial growth (indicating resistance) or inhibition of growth (indicating susceptibility) relative to these controls is determined for each antimicrobial agent-concentration pair. The MIC is defined as the lowest concentration of antimicrobial agent that inhibits growth. Categorical interpretation (SIR result) is furnished based on current FDA or FDA-recognized CLSI breakpoints for each antimicrobial. Species identification by an FDA-cleared test method may be entered at any time during the AST run or after the AST run.

The VITEK® REVEAL™ AST System includes a VITEK® REVEAL™ Sealer, a VITEK® REVEAL™ Instrument, and a master controller computer (MCC)/touch screen monitor. The system is scalable, and up to eight (8) VITEK® REVEAL™ Instruments can be controlled by one user-friendly, touchscreen interface. The VITEK® REVEAL™ AST System is also modular, avoiding the risk of a single instrument failure causing an interruption in laboratory testing. Each VITEK® REVEAL™ Instrument has two independently loadable drawers with each drawer able to hold two (2) GN AST Assays. A single VITEK® REVEAL™ Sealer can support multiple VITEK® REVEAL™ instruments since each sealing step takes less than a minute with a one-button operation.

The provided text describes the VITEK® REVEAL™ GN AST Assay and VITEK® REVEAL™ AST System, an automated system for antimicrobial susceptibility testing. The document details several performance studies to demonstrate the device's accuracy and robustness.

Here's an analysis of the acceptance criteria and study proving the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance

The document presents acceptance criteria for Essential Agreement (EA) and Categorical Agreement (CA) for various performance studies. The overall acceptance criterion for EA and CA in most studies (blood culture bottle equivalency, sample stability, interfering substances) is >89.9% or ≥95%, while for reproducibility, it's ≥95% (best-case) and ≥89% (worst-case).

The table below summarizes the reported performance against these criteria for the Method Comparison Study, which evaluates clinical performance. The full table for all antimicrobial agents and species is extensive, so a representative excerpt from the "Combined" results (Clinical + Challenge samples) for key performance metrics is provided from Table 8 in the document.

Acceptance Criteria for Method Comparison Study (Clinical Performance), as per FDA guidance (Class II Special Controls Document: Antimicrobial Susceptibility Test (AST) Systems):

• Essential Agreement (EA) and Categorical Agreement (CA): Not explicitly stated as a single number at an overall level, but generally expected to be high (e.g., >89.9% for individual errors to be acceptable). The document states "Overall agreement was high".
• Error Rates (Very Major (VMJ), Major (MJ), Minor (MIN)):
  • VMJ (False Susceptible): Should be ≤1.5%.
  • MJ (False Resistant): Should be ≤3.0%.
  • MIN (Minor Error): Not explicitly given a threshold, but typically acceptable if EA is high and clinical impact is low.

Table: VITEK® REVEAL™ Reported Device Performance (Excerpt from Method Comparison, Combined Samples)

Note: The error rates (VMJ, MJ, MIN) were calculated based on the # vmj, # maj, # min columns and Total samples from Table 8, then converted to percentages. Some "N/A" for minors indicate that the count was not provided separately if EA/CA met expectations. The acceptance of lower CA/EA for some combinations is discussed in the text, usually with a rationale (e.g., "all minor errors") and/or a specific limitation added to the product labeling, indicating that the overall performance is considered acceptable by the FDA for market clearance.

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

For the Method Comparison Study (Clinical Performance), which serves as the primary test set:

• Total Samples Enrolled: 1239 samples.
• Total Samples Included in Final Performance Analyses: 1115 samples.
  • Fresh Prospective Positive Blood Culture Samples: 424 samples (after exclusions). These are deidentified clinical samples from patients suspected of bacteremia.
  • Contrived Samples with Clinical Stock Isolates: 101 samples (after exclusions).
  • Contrived Samples with Challenge Isolates: 590 samples (after exclusions).
• Data Provenance:
  • Country of Origin: Not explicitly stated, but "seven (7) US clinical sites" indicates the data is from the United States.
  • Retrospective or Prospective: Both. "Fresh prospective positive blood culture samples" refers to prospectively collected samples. "Clinical stock isolates from the site's inventory and provided challenge isolates" refer to contrived samples, which could be based on retrospective collections or reference strains.

For Reproducibility Study:

• A "set of Gram-negative isolates" was selected. For each panel organism, testing was performed at three sites, in triplicate, on three days, for a total of 27 results per sample. The total number of tests varied per antibiotic, e.g., Amikacin had 891 tests, Amoxicillin/clavulanate had 324 tests etc.

For Blood Culture Bottle Equivalency Study:

• Sample Size: Nine (9) strains representing seven (7) species were tested. Six (6) bottle replicates were tested for each organism and bottle type.

For Sample Stability Study:

• Sample Size: Twenty-five (25) strains representing nine (9) species were selected. Testing was performed in triplicate for each temperature condition and timepoint.

For Interfering Substances Study:

• Sample Size: Five (5) organisms were tested in triplicate for each of the nine (9) interfering substances. This sums to 15 tests per interferant (5 organisms * 3 replicates), which is summed up per interferant type in Table 5a. Additional testing with platelets involved "Fifteen-sixteen strains".

For Initial Inoculum Study:

• Sample Size: Five (5) species were tested in triplicate blood culture bottles for each of three (3) different initial starting concentrations. This means 15 tests per species (5 species * 3 concentrations * 3 replicates = 45 total tests).

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

The document does not explicitly state the "number of experts" or their specific "qualifications" involved in establishing the ground truth.

However, the ground truth for Antimicrobial Susceptibility Testing (AST) results, which is a key component of the device's performance, was established by reference frozen broth microdilution (BMD). This method was "tested according to CLSI M07 (11th Edition) Standard" and "performed in triplicate on custom, 96-well, frozen microdilution plates prepared by the reference testing laboratory." This implies that the ground truth was established by established laboratory protocols and standards (CLSI) using a reference method, not necessarily by a panel of human experts adjudicating results. Personnel performing these reference tests would be trained laboratory professionals.

4. Adjudication Method for the Test Set

The concept of "adjudication" as typically applied to human experts resolving disagreements (e.g., 2+1 or 3+1) is not applicable here.

For the Method Comparison study, the device's MIC results and categorical interpretations were compared against reference BMD modal MICs. The BMD method itself involved triplicate testing. If there were discrepancies in the triplicate BMD results, a "modal MIC" would be used as the reference. Discrepancies between the device and the reference BMD would then be classified into Essential Agreement (EA), Categorical Agreement (CA), and error rates (VMJ, MJ, MIN), against pre-defined breakpoints.

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

No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done.

This device (VITEK® REVEAL™ AST System) is an automated in vitro diagnostic (IVD) system for antimicrobial susceptibility testing. It determines MICs and categorical interpretations directly from positive blood cultures. It does not involve human readers interpreting images or data alongside AI. Therefore, a study comparing human readers with and without AI assistance is not relevant to this type of device. The study evaluates the device's standalone performance against a gold standard laboratory method (BMD).

6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done

Yes, a standalone performance evaluation was done.

The entire performance study, particularly the "Method Comparison Study," assesses the VITEK® REVEAL™ AST System's ability to accurately determine antimicrobial susceptibility on its own, by comparing its results directly to the reference BMD method. The system is designed to provide automated results without human interpretation or intervention in the determination of susceptibility profiles. The human role is in sample preparation and inputting organism identification (obtained from another FDA-cleared rapid ID method), but the AST determination itself is performed by the system's algorithm and sensors.

7. The Type of Ground Truth Used

The ground truth used for the performance studies, especially the critical Method Comparison Study, was:

• Reference frozen broth microdilution (BMD), tested according to CLSI M07 (11th Edition) Standard. This is a widely accepted laboratory gold standard for antimicrobial susceptibility testing.

Additionally, for the Method Comparison, "Organism identification for all samples was confirmed by an FDA-cleared MALDI ID method." This serves as ground truth for the organism identity, which is a prerequisite for accurate AST interpretation.

8. The Sample Size for the Training Set

The document does not provide information about a "training set" or its size. This is typical for in vitro diagnostic devices like the VITEK® REVEAL™ System, especially those based on biophysical detection rather than machine learning algorithms trained on large datasets. The primary development and validation for such systems often involve engineering principles, analytical studies, and direct comparisons to reference methods, rather than distinct "training" and "test" sets in the machine learning sense. The performance data presented (reproducibility, equivalency, stability, interfering substances, and method comparison) are analogous to what would be considered validation or test data to prove the device's capability.

If there are internal algorithms that "learn" or optimize, the training data for such components and their size are not disclosed in this document.

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

As noted in point 8, information about a "training set" and its ground truth establishment is not available in the provided document. The development of the VITEK® REVEAL™ System's proprietary chemical Small Molecule Sensors (SMS) and its real-time algorithm for detecting growth, as described in the device description section, would have involved significant internal research and development, but the specific validation process for these "training" aspects is not detailed in this 510(k) summary. The document focuses on the performance validation for regulatory submission.

Ask a specific question about this device

Intended Use:
The ASTar System is intended to be used for the automated quantitative susceptibility testing for most clinically significant microorganisms. The ASTar System does not provide organism identification.

Indications for Use:
The ASTar System, comprised of the ASTar Instrument with the ASTar BC G- Kit (ASTar BC G- Consumable kit, ASTar BC G- Frozen insert, and ASTar BC G-Kit software), utilizes high-speed, time-lapse microscopy imaging of bacteria for the in vitro, quantitative determination of antimicrobial susceptibility of on-panel gram-negative bacteria. The test is performed directly on positive blood culture samples signaled as positive by a continuous monitoring blood culture system and confirmed to contain gram-negative bacilli by Gram stain. Organism identification is required for AST result interpretation and reporting.

Test results from the ASTar BC G- Kit should be interpreted in conjunction with other clinical and laboratory findings. Standard laboratory protocols for processing positive blood cultures should be followed to ensure availability of isolates for supplemental testing. Sub-culturing is necessary to support further testing for: bacteria and antimicrobials not on the ASTar BC G- panel, where inconclusive results are obtained, epidemiologic testing, recovery of organisms present in microbial samples, and susceptibility testing of bacteria in polymicrobial samples.

The ASTar BC G- Kit tests the following antimicrobial agents with the following bacterial species:

Amikacin: Citrobacter freundii, Enterobacter cloacae complex, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Pseudomonas aeruginosa, Serratia marcescens
Ampicillin: Escherichia coli, Proteus mirabilis
Ampicillin-sulbactam: Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris
Aztreonam: Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Serratia marcescens
Cefazolin: Klebsiella pneumoniae
Cefepime: Citrobacter freundii, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens
Ceftazidime: Enterobacter cloacae complex, Escherichia oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Serratia marcescens
Ceftazidime-avibactam: Citrobacter freundii, Citrobacter koseri, Enterobacter cloacae complex, Klebsiella oxytoca, Proteus mirabilis, Pseudomonas aeruginosa, Serratia marcescens
Cefuroxime: Escherichia coli, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis
Ciprofloxacin: Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens
Gentamicin: Citrobacter freundii, Citrobacter koseri, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens
Levofloxacin: Citrobacter freundii, Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens
Meropenem: Acinetobacter baumannii, Citrobacter freundii, Citrobacter koseri, Escherichia coli, Proteus mirabilis, Proteus vulgaris, Pseudomonas aeruginosa, Serratia marcescens
Meropenem-vaborbactam: Citrobacter freundii, Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella pneumoniae, Proteus mirabilis, Serratia marcescens
Piperacillin-tazobactam: Citrobacter koseri, Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, Proteus vulgaris, Serratia marcescens
Tigecycline: Citrobacter freundii, Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Serratia marcescens
Tobramycin: Citrobacter freundii, Citrobacter koseri, Enterobacter cloacae complex, Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, Serratia marcescens
Trimethoprim-sulfamethoxazole: Enterobacter cloacae complex, Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella pneumoniae, Proteus vulgaris

ASTar System is a fully automated system for antimicrobial susceptibility testing (AST). It consists of the ASTar Instrument which is used in combination with dedicated application kits. The ASTar BC G- Kit consists of the ASTar BC G- Consumable kit, ASTar BC G- Frozen insert, and ASTar BC G-Kit software which must be installed on the instrument to process the kit.

The system provides robust and consistent inoculum preparation for AST and utilizes high-speed, time-lapse microscopy imaging of pathogens in broth microdilution to determine minimum inhibitory concentration (MIC) and qualitative susceptibility results. Organism identification using an approved method is required to be entered into the ASTar Instrument for results to be reported.

The instrument is designed to carry out sample preparation of up to six samples in parallel, using a dedicated ASTar Cartridge consumable for each sample. In the subsequent AST culturing step, the instrument transfers the prepared sample into a second dedicated consumable, referred to as the ASTar Disc. Up to 12 Discs can be incubated simultaneously in the system. The processed samples can be in different stages of the processing protocol. New samples can be loaded in a random-access manner when there are available slots. Processing of loaded samples will, in most cases, start shortly after loading. If six samples are started at the same time limitations given by the sample scheduler will result in a queue. The operator interacts with the instrument via the touchscreen display by which the operator controls the instrument.

ASTar BC G- Kit is used for in vitro determination of antimicrobial susceptibility testing of commonly isolated bacteria derived from positive blood culture samples confirmed positive for Gram-negative bacteria by Gram stain. The antimicrobial and organism combinations are listed in Table 1. Reportable ranges for each antimicrobial are listed in Table 2.

To start an analysis approximately 1 mL of a positive blood culture, confirmed Gram-negative by Gram stain is pipetted into the ASTar Cartridge by the operator and loaded into the system, from which the system purifies and quantifies the bacterial concentration is adjusted to the appropriate inoculum concentration and produces an inoculum for analysis of non-fastidious organisms. The bacterial suspensions are transferred automatically to the ASTar Disc and antimicrobial susceptibility testing is performed based on a defined short-term protocol. Results are available within approximately six hours. Bacterial growth and response to relevant concentrations of different antimicrobial drugs are measured throughout the incubation period, using a high-performance optical detection system in combination with image analysis algorithms. The system generates an MIC and further qualitative susceptibility results (i.e., S, I, R) for the tested antimicrobials when applicable. The qualitative results are determined based on established breakpoints stipulated by applicable authorities, i.e., FDA, CLSI or EUCAST. FDA Susceptibility Testing Interpretive Criteria (STIC), aka "breakpoints" are found in Table 3.

The provided text describes the performance characteristics of the ASTar BC G- Kit and ASTar Instrument, primarily focusing on its antimicrobial susceptibility testing (AST) capabilities. While it details various studies, it does not describe an AI/ML device that utilizes a test set with ground truth experts. Instead, it describes a medical device for in vitro quantitative determination of antimicrobial susceptibility based on time-lapse microscopy imaging.

Therefore, many of the requested points, such as "number of experts used to establish ground truth," "adjudication method," "MRMC comparative effectiveness study," "standalone (algorithm only) performance," and "sample size for the training set" (for an AI model), are not applicable to this document as it does not describe an AI/ML-driven diagnostic device in the traditional sense.

However, I will extract relevant information about the device's acceptance criteria and studies to the best of my ability, interpreting "acceptance criteria" as performance metrics for this type of medical device.

Key Information from the Document:

The ASTar System is an automated system for antimicrobial susceptibility testing (AST) that uses high-speed, time-lapse microscopy imaging of bacteria to determine Minimum Inhibitory Concentration (MIC) and qualitative susceptibility results (S, I, R).

1. A table of acceptance criteria and the reported device performance

The document defines acceptance criteria primarily through performance metrics like Essential Agreement (EA) and Category Agreement (CA) compared to a reference method (frozen Broth Micro-Dilution, BMD), along with rates for Very Major (VMJ) discordant results, Major (MAJ) discordant results, and Minor (MIN) discordant results.

While a single explicit "acceptance criteria table" is not provided with specific pass/fail percentages before results, the overall performance table (Table 16) implicitly represents the success or failure against internal performance goals. The FDA's Special Controls guidance (referenced in 8.5.8) would typically outline such criteria. Based on the "Conclusions" section, the device was deemed "substantially equivalent," implying these metrics were acceptable.

Here's a summary of the reported device performance from Table 16, which reflects the met acceptance criteria for the clinical study:

Table: Reported Device Performance (Summary from Table 16)

*Note: Some "poor performance" combinations (EA 95% of MIC values within ±1 doubling dilution of the mode MIC of initial samples (loaded 95% pass rate as compared to control samples without interfering antibiotics.

Interfering Antibiotics Performance (Table 13):

• All six evaluated antibiotic/BCB-combinations had overall pass rates of 96.2% to 100%. Some individual combinations fell below 90% (e.g., Cefotaxime / BACTEC: Trimethoprim-sulfamethoxazole 77.8%), but the overall criterion (per combination type) was met.

Carry Over and Cross Contamination Acceptance Criteria: (Implicitly, close to 100% pass rate expected)

• MIC for the susceptible isolate for each antimicrobial must be within ±1 doubling dilution of the control mode MIC to pass.

Carry Over and Cross Contamination Performance:

• 99.7% pass rate (307/308) for susceptible isolate MIC value. No carry over or cross contamination observed.

Set Inoculum for AST Acceptance Criteria:

• For starting bacterial concentration >5 x 10^7 CFU/mL, assess and adjust successfully at high rate, producing an inoculum within acceptance ranges.
• For concentrations 5 x 10^7 CFU/mL: 95.8% (23/24) completed concentration adjustment, and 100% (23/23) of those produced an inoculum within acceptance ranges.
• For samples with starting bacterial concentration 5 x 10^6 to

Ask a specific question about this device

The LifeScale AST system is a multiplexed in vitro diagnostic test that uses and resonant frequency to calculate organism concentration and/or mass distribution for quantitative antimicrobial susceptibility testing is performed directly on blood cultures signaled as positive by a continuous monitoring blood culture system and confirmed by Gram stain. The LifeScale AST system does not provide organism identification and is not indicated for use with polymicrobial samples. Interpretive results (Susceptible/Intermediate/Resistant) are provided for specific drug/organism combinations. Results are used in conjunction with other clinical and laboratory findings. Standard laboratory protocols for processing positive blood cultures should be followed to ensure availability of isolates for supplemental testing as needed. Additionally, subculture is necessary for the susceptibility testing of organisms present in polymicrobial samples, for testing antimicrobial agents and species not indicated for testing with the device, for epidemiologic testing and for recovery of organisms present in microbial samples.

The LifeScale Gram Negative Kit (LSGN) is intended for use with the LifeScale AST system for in vitro testing of positive blood culture samples confirmed by Gram staining gram-negative bacilli for the antimicrobial agents and specific target organisms identified below:

• Ampicillin: Escherichia coli

• Aztreonam: Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca
• Cefazolin: Klebsiella pneumoniae, Klebsiella variicola
• Ceftazidime: Acinetobacter baumannii/hosocomialis group, Escherichia coli, Klebsiella aerogenes,
• Klebsiella oxytoca, Klebsiella variicola, Pseudomonas aeruginosa
• · Ertapenem: Escherichia coli, Klebsiella aerogenes, Klebsiella oxytoca
• Trimethoprim-Sulfamethoxazole: Escherichia coli, Klebsiella oxytoca, Klebsiella oxytoca, Klebsiella variicola

The LifeScale AST system is an in vitro diagnostic test designed to quantitatively assess antimicrobial susceptibility using a microfluidic sensor and resonant frequency technology. Specifically engineered for use with positive blood culture samples confirmed positive by Gram stain for Gram-negative rods, the LifeScale LSGN Panel ensures compatibility and accuracy while excluding Gram-positive or polymicrobial samples, thus maintaining specificity and reliability. During the incubation phase, the LifeScale LSGN Panel offers a standard incubation time of 3 hours, extendable up to 6 hours to accommodate varying microbial growth rates. Panels must be read within 8 hours of setup, with automatic cancellation for panels exceeding this timeframe. Panels with delayed readings can be safely stored in the offline incubator until analysis. Upon reaching sufficient growth in the positive control wells, the LifeScale AST system transitions to data acquisition and readout. Advanced sensors capture essential metrics including microbe count, mass, and fluid volume, processed through sophisticated software algorithms to generate precise AST results for each antibiotic. To maintain hygiene standards, the LifeScale AST system incorporates automated washing and disinfection protocols for the sipper and sensor, minimizing the risk of cross-contamination and organic buildup. The culmination of the testing process involves calculating and reporting AST results (MIC and interpretive results), providing clinicians with actionable insights into antibiotic efficacy. Species-level organism identification is essential for results reporting. AST results are generated based on FDA or CLSI breakpoints validated for laboratory use.

Acceptance Criteria and Device Performance Study for LifeScale Gram Negative Kit (LSGN) with the LifeScale AST system (K211815)

The information provided describes the performance of the LifeScale Gram Negative Kit (LSGN) with the LifeScale AST system in comparison to a reference method (Broth Microdilution Method - BMD) for Antimicrobial Susceptibility Testing (AST) of gram-negative bacilli from positive blood cultures.

1. Table of Acceptance Criteria and Reported Device Performance

The general acceptance criteria for clinical performance for each antimicrobial agent on the LSGN panel were >90% Essential Agreement (EA) and >90% Categorical Agreement (CA) rates. Additionally, specific thresholds for Very Major Discrepancy (VMJ), Major Discrepancy (MAJ), and Minor Discrepancy (MIN) were evaluated, but the primary overall acceptance for accuracy relies on EA and CA.

Overall Clinical Performance (Aggregated across organisms for each drug)

*For Ceftazidime (P. aeruginosa), major errors were 8.20% (5/61 susceptible isolates), adjusted to 3.3% (2 major errors) due to the lack of an intermediate breakpoint. This value is flagged as a limitation requiring an alternative testing method.
^For Ertapenem (K. oxytoca), very major errors were 20.0% (2/10 resistant isolates). This value is flagged as a limitation requiring an alternative testing method.

Reproducibility Acceptance Criteria and Performance

Blood Bottle Compatibility Acceptance Criteria and Performance
Acceptance Criteria: ≥90% EA for various blood culture media compared to LifeScale LSGN and Reference BMD modes. All tested media types met this criterion, except for specific drug/organism combinations that resulted in limitations (e.g., ertapenem/E.coli with certain anaerobic media).

Sample Stability Acceptance Criteria and Performance
Acceptance Criteria: ≥90% agreement with the reference method modal MIC and the LifeScale LSGN MIC mode within +/- one two-fold dilution at Tpos and T13. All listed antimicrobial/organism combinations met this criterion.

Interfering Substances Acceptance Criteria and Performance
Acceptance Criteria: EA to the mode of control MICs ≥ 90% for all interfering substances. This criterion was generally met for all substances and antibiotics, with some discrepancies noted but attributed to individual strain behavior rather than the interfering substance.

Inoculum Density Study Acceptance Criteria and Performance
Acceptance Criteria: ≥90% essential agreement (within +/- 1 antibiotic dilution) compared to the LifeScale LSGN mode. All tested antimicrobial/organism combinations at 10e6 and 10e9 CFU/mL met this criterion.

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

The study included both prospective clinical blood cultures (PBCs) and contrived samples, as well as CDC Challenge and other reference laboratory isolates.

• Clinical Performance Data:

  • Total number of samples evaluated varied by antimicrobial agent:
    • Ampicillin: 137 samples
    • Aztreonam: 301 samples
    • Cefazolin: 143 samples
    • Ceftazidime: 529 samples (total across species)
    • Ertapenem: 224 samples
    • Trimethoprim-Sulfamethoxazole: 340 samples
  • Data Provenance:
    • Prospective clinical blood cultures (PBCs): Enrolled and tested at 6 US Clinical sites. These were verified by Gram stain for gram-negative bacilli.
    • Contrived samples: Prepared from frozen isolates supplied by Affinity Biosensors or from contemporary isolates collected by the laboratory. Blood cultures were spiked and incubated.
    • CDC Challenge and Challenge isolates: Prepared from frozen isolates supplied by Affinity Biosensors to one trial site.
  • The split between clinical (prospective and seeded) and challenge samples varied by antimicrobial (e.g., Ampicillin: 91 clinical, 46 challenge; Aztreonam: 235 clinical, 66 challenge).

• Reproducibility Study: Total evaluable results varied per antibiotic (e.g., Ampicillin: 269; Aztreonam: 294). This involved testing at three testing sites over three consecutive days with three replicates per day for each test organism (total 27 data points per organism).

• Blood Bottle Compatibility: A minimum of 12 strains were tested using each blood culture media type, with replicates of ten (10) per organism/media type. Organisms used were Escherichia coli (8 strains), Klebsiella pneumoniae (4 strains), Pseudomonas aeruginosa (2 strains), Acinetobacter baumannii (3 strains).

• Sample Stability: 16 organisms (susceptible and resistant on-scale MICs for each relevant antibiotic) were chosen. Each organism was tested in triplicate on three LifeScale AST systems at two time points (Tpos and T13). The study evaluated 126 LifeScale LSGN samples for total performance.

• Interfering Substances: A minimum of five (5) test organisms were used for each interferent. Each interfering substance was tested in triplicate for each test organism/media combination. Organisms included Escherichia coli (3 strains), Pseudomonas aeruginosa (1 strain), Acinetobacter baumannii (1 strain), Klebsiella pneumoniae (2 strains).

• Inoculum Density Study: Thirteen strains (not specified if unique per antibiotic) were tested at target organism concentrations of 10e6 CFU/ml and 10e9 CFU/ml. Slow-growing organisms were also seeded at 10e4 CFU/ml.

• Contamination and Carry-Over Testing:

  • Test 1 (Plate-to-Plate with Growth Media): 5 pairs of plates. Each pair read on a different LifeScale AST system.
  • Test 2 (Plate-to-plate with Resistant and Susceptible Organisms): 2 organisms (K. pneumoniae AR0107 and E. coli ATCC25922). Paired plates tested in triplicate and read on different LifeScale AST systems.
  • Test 3 (Well-to-Well Cross Carry-Over): 5 plates. Each plate read on a different LifeScale AST system.

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

The document does not explicitly state the number or qualifications of experts used to establish the ground truth for the test set. However, it does mention that "Reference testing was performed on all enrolled samples in triplicate. Testing was done in accordance with the reference protocol and was performed at one clinical site shipped isolates on transport media from the PBC purity panel following verification of pure culture." This implies that trained laboratory personnel performed the reference testing, but their specific qualifications as "experts" for ground truth establishment are not detailed.

4. Adjudication Method for the Test Set

The document states that "Each sample submitted for BMD testing was assigned a unique Trial ID, and LifeScale results were kept blinded to prevent bias. Performance was evaluated by comparing quantitative (MIC) and qualitative (S-I-R) AST results generated by the LifeScale LSGN kit with those of the reference BMD." This describes a blinded comparison method against a reference standard (BMD). There is no explicit mention of an adjudication panel (e.g., 2+1, 3+1) for discordant results.

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 is not an MRMC comparative effectiveness study involving human readers or AI assistance in interpretation. The device is an automated in vitro diagnostic test for antimicrobial susceptibility. Its performance is compared directly against a reference laboratory method (Broth Microdilution Method), not against human interpretation or for human improvement with AI assistance. Therefore, this section is not applicable.

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

Yes, a standalone performance evaluation was done. The LifeScale AST system is described as an automated microfluidic sensor system that uses resonant frequency and sophisticated software algorithms to generate AST results (MIC and interpretive results). The entire performance evaluation described focuses on the device's ability to accurately produce these results compared to the reference BMD method without human intervention in the result generation process itself. Human input is for organism ID (from external systems) for final reporting, but the core AST determination is standalone.

7. The Type of Ground Truth Used (expert consensus, pathology, outcomes data, etc.)

The primary ground truth used for assessing the performance of the LifeScale LSGN kit was the Broth Microdilution Method (BMD), in accordance with CLSI guidance (CLSI M07). This is considered the gold standard reference method for antimicrobial susceptibility testing in microbiology.

8. The Sample Size for the Training Set

The document does not provide details on the sample size for the training set for the LifeScale AST system's algorithms. The studies described are for the performance evaluation of the final device using a defined test set.

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

As no information regarding a separate training set is provided, no details are available on how the ground truth for any potential training set was established. It is common for such systems to be developed/trained using a combination of laboratory-contrived samples with known AST profiles (established by methods like BMD) and potentially retrospective clinical samples. However, the document does not elaborate on this aspect.

Ask a specific question about this device