Search Results

The Sensititre® 18 - 24 hour MIC or Breakpoint Susceptibility System is an in vitro diagnostic product for clinical susceptibility testing of non-fastidious isolates. This 510(k) is for the removal of the limitation for the ability to detect resistance for Meropenem (0.004 - 8μg/mL) and Enterobacteriaceae and for the addition of the newly approved breakpoints (S≤1, |=2, R≥4) on the Sensititre® 18 - 24 hour MIC panel for testing Gram negative isolates. The approved primary, "Indications for Use" and clinical significance for Enterobacteriaceae is for the following species: Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis. In vitro data, without clinical correlation is provided for: Aeromonas hydrophila, Citrobacter koseri (formerly diversus), Citrobacter freundii, Enterobacter cloacae, Hafnia alvei, Klebsiella oxytoca, Morganella morganii, Proteus vulgaris, Serratia marcescens.

Not Found

This document is an FDA 510(k) clearance letter for a medical device called the "Sensititre 18-24 hour MIC Susceptibility System." As such, it does not contain the detailed study information generally associated with acceptance criteria and device performance studies (e.g., sample sizes, expert qualifications, ground truth methodology) for a typical AI/ML-based medical device.

Instead, this document focuses on the regulatory clearance process for an in vitro diagnostic product (antimicrobial susceptibility test). The "acceptance criteria" here refer to the performance standards for antimicrobial susceptibility testing (AST), which are typically defined by regulatory bodies (like FDA) and standardized by organizations (like CLSI - Clinical and Laboratory Standards Institute).

The key performance metrics for AST devices focus on how accurately they categorize bacterial isolates as susceptible, intermediate, or resistant to an antibiotic.

However, based on the provided text, I can extract and infer some information, particularly regarding the Indications for Use and the context of the device clearance.

Here's a breakdown of what can be extracted/inferred:

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

The document doesn't provide a typical table of statistical acceptance criteria (e.g., sensitivity, specificity thresholds) and corresponding reported performance values like a clinical study report would. Instead, it states the purpose of the 510(k) is:

• "for the removal of the limitation for the ability to detect resistance for Meropenem (0.004 - 8μg/mL) and Enterobacteriaceae"
• "for the addition of the newly approved breakpoints (S≤1, I=2, R≥4) on the Sensititre® 18 - 24 hour MIC panel for testing Gram negative isolates."

This implies that the "acceptance criteria" were met by demonstrating that the device:

1. Can reliably detect resistance to Meropenem for Enterobacteriaceae within the specified MIC range.
2. Accurately applies the updated breakpoints for Meropenem in Gram-negative isolates.

To meet these, the device's performance would have been compared against a reference method (likely broth microdilution or agar dilution as per CLSI standards) and statistical agreement measured. The FDA would have assessed Essential Agreement (EA) and Categorical Agreement (CA), and typically very major, major, and minor error rates. While the specific numerical acceptance criteria (e.g., >90% EA, >90% CA) commonly used for ASTs are not explicitly stated in this letter, they would have been part of the underlying submission.

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective):

• Sample Size: Not specified in the provided document. For AST devices, sample sizes are typically determined by the number of unique clinical isolates tested, often across multiple strains and resistance mechanisms, as per CLSI guidelines, to ensure adequate statistical power for assessing agreement.
• Data Provenance: Not specified in the provided document. Such studies are generally conducted in CLSI-compliant laboratories, often in the US, but the specific country of origin is not mentioned. They would be prospective in the sense that the device is being tested on newly identified bacterial strains, though the strains themselves might be from a collection (retrospective repository) but cultured and tested prospectively.

3. 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):

• This type of device (antimicrobial susceptibility test) does not typically involve human "experts" establishing a "ground truth" in the way an imaging AI device would.
• Ground Truth Establishment: The ground truth for antimicrobial susceptibility testing is established by a reference method, most commonly a standardized broth microdilution or agar dilution method, as defined by organizations like CLSI. The interpretation of these reference method results (e.g., determining S, I, or R based on MIC values) follows established guidelines, not individual expert opinion. The "experts" involved would be trained microbiologists performing the reference method in a GLP-compliant laboratory environment.

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

• Adjudication Method: Not applicable for this type of device. Adjudication methods (like 2+1 or 3+1) are used to resolve disagreements between human readers in image interpretation. Here, the ground truth is an objective laboratory reading, and the device's result is compared directly to it.

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:

• MRMC Study: Not applicable. This is an automated in vitro diagnostic device, not an AI-assisted diagnostic tool designed to improve human reader performance. The device provides a result directly.

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

• Standalone Performance: Yes, this device's performance is inherently "standalone." It is an automated system that determines antimicrobial susceptibility without human interpretation of raw data points to categorize. Human involvement is in setting up the test and interpreting the final device output.

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

• Ground Truth: The ground truth for this device is established by a standardized reference method, typically broth microdilution or agar dilution, interpreted according to established clinical breakpoints (e.g., CLSI guidelines).

8. The sample size for the training set:

• Training Set: Not applicable in the context of traditional machine learning training sets for AI devices. This device is likely based on chemical reactions and optical readings rather than a deep learning algorithm trained on a large dataset of results. Its "development" phase would involve optimizing the reagents, reading algorithms, and interpretive logic.

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

• Training Set Ground Truth: Not applicable, as explained above. The underlying principles are established microbiological methods and their correlation to clinical outcomes over decades of research, rather than a machine learning training process with a specific ground truth dataset.

In summary, the provided document is a regulatory clearance for a traditional in vitro diagnostic device, not an AI/ML-based device. Therefore, many of the requested details about study design (e.g., expert-driven ground truth, MRMC studies, AI training sets) are not relevant or present in this type of document. The acceptance criteria would have been met by demonstrating sufficient agreement (essential and categorical) with a recognized reference method for antimicrobial susceptibility testing.

Ask a specific question about this device