(1026 days)
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)
| Antimicrobial Agent | Acceptance Criteria (EA%) | Reported EA% | Acceptance Criteria (CA%) | Reported CA% | Acceptance Criteria (VMJ%) (Max) | Reported VMJ% | Acceptance Criteria (MAJ%) (Max) | Reported MAJ% |
|---|---|---|---|---|---|---|---|---|
| Ampicillin | >90% | 100.0% | >90% | 100.0% | Not explicit, implied low | 0.00% | Not explicit, implied low | 0.00% |
| Aztreonam | >90% | 98.0% | >90% | 98.0% | Not explicit, implied low | 0.00% | Not explicit, implied low | 0.48% |
| Cefazolin | >90% | 97.9% | >90% | 92.3% | Not explicit, implied low | 0.00% | Not explicit, implied low | 0.00% |
| Ceftazidime (E. coli, K. aerogenes, K. oxytoca, K. variicola) | >90% | 97.6% | >90% | 98.2% | Not explicit, implied low | 0.00% | Not explicit, implied low | 0.00% |
| Ceftazidime (P. aeruginosa) | >90% | 92.2% | >90% | 94.0% | Not explicit, implied low | 3.64% | Not explicit, implied low | 8.20%* |
| Ceftazidime (A. baumannii, A. baumannii/nosocomialis group) | >90% | 98.6% | >90% | 100.0% | Not explicit, implied low | 0.00% | Not explicit, implied low | 0.00% |
| Ertapenem | >90% | 93.8% | >90% | 95.6% | Not explicit, implied low | 4.65%^ | Not explicit, implied low | 0.55% |
| Trimethoprim-Sulfamethoxazole | >90% | 99.1% | >90% | 99.1% | Not explicit, implied low | 0.00% | Not explicit, implied low | 1.20% |
*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
| Antibiotic | Acceptance Criteria (Best-case %) | Reported Best-case (%) | Acceptance Criteria (Worst-case %) | Reported Worst-case (%) |
|---|---|---|---|---|
| Ampicillin | >95% | 99.6% | >89% | 92.9% |
| Aztreonam | >95% | 98.6% | >89% | 89.5% |
| Cefazolin | >95% | 99.1% | >89% | 99.1% |
| Ceftazidime | >95% | 99.6% | >89% | 92.1% |
| Ertapenem | >95% | 98.3% | >89% | 91.6% |
| Trimethoprim-Sulfamethoxazole | >95% | 96.6% | >89% | 92.0% |
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).
- Total number of samples evaluated varied by antimicrobial agent:
-
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.
§ 866.1650 A cellular analysis system for multiplexed antimicrobial susceptibility testing.
(a)
Identification. A cellular analysis system for multiplexed antimicrobial susceptibility testing is a multiplex qualitative and/or quantitative in vitro diagnostic device intended for the identification and determination of the antimicrobial susceptibility results of organisms detected in samples from patients with suspected microbial infections. This device is intended to aid in the determination of antimicrobial susceptibility or resistance when used in conjunction with other laboratory findings.(b)
Classification. Class II (special controls). The special controls for this device are:(1) Design verification and validation must include:
(i) Detailed device description documentation, including the device components, ancillary reagents required but not provided, a detailed explanation of the methodology, including primer/probe sequence, design, rationale for sequence selection, and details of the antimicrobial agents, as applicable.
(ii) Detailed documentation from the following analytical and clinical performance studies: limit of detection, inclusivity, precision, reproducibility, interference, cross-reactivity, carryover, and cross-contamination, quality control and additional studies, as applicable to specimen type and assay intended use.
(iii) Detailed documentation from an appropriate clinical study. The study, performed on a study population consistent with the intended use population, must compare the device performance to results obtained from well-accepted reference methods.
(iv) Detailed documentation for device software, including software applications and hardware-based devices that incorporate software.
(2) The labeling required under § 809.10(b) of this chapter must include:
(i) Limitations and protocols regarding the need for correlation of results by standard laboratory procedures, as applicable.
(ii) A detailed explanation of the interpretation of results and acceptance criteria.
(iii) A detailed explanation of the principles of operation and procedures for assay performance and troubleshooting.