Search Results

The provided document is a 510(k) clearance letter from the FDA for the Ambu Neonate Silicone Resuscitator. It does not contain the acceptance criteria or details of a study that proves the device meets specific acceptance criteria in the way a performance study for an AI/ML device would.

Instead, this document grants clearance based on the determination that the device is substantially equivalent to legally marketed predicate devices. This means that the FDA reviewed information submitted by Ambu (in their 510(k) submission, not provided here) and concluded that the new device is as safe and effective as existing devices on the market for its intended use.

Therefore, I cannot populate the requested table or answer the specific questions about sample sizes, ground truth, expert qualifications, or MRMC studies for this particular document, as those details are not part of a 510(k) clearance letter. These types of studies and criteria are typically associated with performance testing for new and complex medical devices, especially AI/ML-driven ones, and are included in the underlying 510(k) submission, not the clearance letter itself.

The key information from this document is:

Device Name: Ambu Neonate Silicone Resuscitator
Regulatory Class: II (two)
Product Code: 73 BTM
Indications For Use: The Ambu Neonate Silicone Resuscitator is intended for ventilation of neonates and infants with a body weight up to 22 lb (10 kg).
Approval Date: May 17, 2000
Basis for Approval: Substantial equivalence to legally marketed predicate devices.

If this were a description of an AI-driven device and its performance study, the information would be structured as follows:

(Hypothetical Example if the document did contain the requested information for an AI device)

1. Table of Acceptance Criteria and Reported Device Performance

Performance Metric	Acceptance Criteria	Reported Device Performance
Primary Endpoints
[Metric 1, e.g., Sensitivity for X]	> [Value, e.g., 90%] with 95% CI	[Reported Value, e.g., 92.5% (95% CI: 90.1-94.2%)]
[Metric 2, e.g., Specificity for Y]	> [Value, e.g., 85%] with 95% CI	[Reported Value, e.g., 86.1% (95% CI: 83.9-88.0%)]
Secondary Endpoints
[Metric 3, e.g., AUC]	> [Value, e.g., 0.90]	[Reported Value, e.g., 0.93]
[Metric 4, e.g., Time to Detection]	10 years of experience in pulmonary imaging, and 2 expert thoracic surgeons.]

4. Adjudication method for the test set:

[e.g., 2+1 (two initial readers, third adjudicator for disagreement), 3+1 (three initial readers, 4th adjudicator for disagreement), Simple majority, Consensus meeting, No adjudication (single reader for ground truth)]

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:

[Yes/No]
Effect Size: [e.g., Human readers' sensitivity for lesion detection improved by an average of X% (e.g., 15%) when using AI assistance (from Baseline Y% to Z%), accompanied by a reduction in reading time by A% (e.g., 20%).]

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

[Yes/No]

7. The type of ground truth used:

[e.g., Expert consensus (as described in #3), Pathology reports (biopsy-confirmed), Clinical outcome data (validated by long-term follow-up), Surgical findings, Combination of the above.]

8. The sample size for the training set:

Sample Size (Training Set): [Number] cases (e.g., 50,000 images)

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

[e.g., Each case was annotated by a single board-certified radiologist, followed by a double-check by a senior radiologist. Automated extraction from electronic health records, where diagnoses were confirmed by at least two independent clinical notes. Utilized a combination of expert labeling and pathology reports for a subset of the data.]

Ask a Question

Ask a specific question about this device

Page 1 of 1