(53 days)
The Gauss Pixel App is indicated for use to aid current practices in recording the number of surgical sponges and for visibility for assessment of sponge images.
The Gauss Pixel App is a software program used on an iPad tablet to capture images of sponges to assist surgical personnel in the management of surgical sponges after surgical The App allows surgical personnel to categorize sponges by sponge type and use. provides an automated ongoing count of total sponge images and sponge images by tag. It also provides a visual record of images for further evaluation. This program is not intended to replace existing sponge counting practices and sponges should be retained per the user's standard sponge management practice until the case is complete and sponge counting has been finalized.
Here's a summary of the acceptance criteria and study findings for the Gauss Pixel App, based on the provided document:
1. Acceptance Criteria and Reported Device Performance
| Acceptance Criteria | Reported Device Performance (Gauss Pixel App) |
|---|---|
| Record Images as Indicated by User | Confirmed: The application provided instructions for use and recorded images as indicated by the user. |
| Accurately Tag Images as Indicated by User | Confirmed: The application accurately tagged images as indicated by the user. |
| Accurately Provide Automated Counting (Total) | Confirmed: The application accurately provided automated counting both in total. |
| Accurately Provide Automated Counting (by Type) | Confirmed: The application accurately provided automated counting by type. |
| Allow Visual Review and Management of Images | Confirmed: The application allowed visual review and management (re-tagging, deletion) of all images as appropriate. |
| Function as Intended | Demonstrated: Results of performance testing through the software verification and validation process demonstrate that the Gauss Pixel App functions as intended. |
| Substantial Equivalence to Predicate Device | Demonstrated: The Gauss Pixel App is as safe and effective as the predicate, has the same intended uses and indications, and utilizes a new technological method (software) which complements current clinical practices and does not raise new issues of safety or effectiveness. Software verification and validation demonstrate it functions as intended. |
2. Sample Size and Data Provenance
The document does not explicitly state the sample size used for a test set in the conventional sense of a clinical trial or independent validation. The performance claims are based on "software verification and validation testing." This typically involves internal testing by the developer to ensure the software meets its specified functional and non-functional requirements.
There is no information provided regarding the country of origin of data or whether it was retrospective or prospective. Given the nature of a 510(k) submission for a Class I device and the focus described (software verification and validation), it's highly probable that this involved internal testing rather than a large-scale clinical study with external patient data.
3. Number of Experts and Qualifications
The document does not specify the number of experts, their qualifications, or their role in establishing a "ground truth" for a test set. The validation described is primarily around the software's functional performance (e.g., counting, tagging, image display) rather than assessing clinical accuracy against expert opinion in a diagnostic context.
4. Adjudication Method
No adjudication method (e.g., 2+1, 3+1) is mentioned, as the described validation focuses on software functionality rather than interpreting findings against a 'ground truth' that would require expert consensus.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No MRMC comparative effectiveness study is mentioned. The device is described as an adjunctive tool to aid current practices, not as a replacement or independent diagnostic tool. The focus is on verifying its own functionality (counting, image display) rather than its impact on human reader performance. Therefore, an effect size of human readers improving with/without AI assistance is not reported.
6. Standalone Performance
Yes, a standalone (algorithm only without human-in-the-loop performance) assessment was done for the core functionalities of the application. The document states: "Results of performance testing confirmed that the application provided instructions for use, recorded images as indicated by the user, accurately tagged images as indicated by the user, accurately provided automated counting both in total and by type and allowed visual review and management (re-tagging, deletion) of all images as appropriate." This describes the independent functional performance of the software.
7. Type of Ground Truth Used
For the described performance testing, the "ground truth" appears to be the expected functional behavior of the software as defined by its specifications and design. For example:
- For "accurately provided automated counting," the ground truth would be the actual number of images taken and the actual type assigned by a user, against which the software's automated count is compared.
- For "accurately tagged images," the ground truth is the tag the user intended to apply, against which the software's recorded tag is compared.
It does not refer to medical ground truth like pathology, outcomes data, or expert consensus on a diagnostic interpretation.
8. Sample Size for the Training Set
The document does not specify a sample size for a training set. This is consistent with the nature of the device as a functional tool for managing sponges, not a machine learning model that requires a large training dataset to develop its core functionality (like image recognition for diagnostic purposes). The "training" described implicitly refers to the software development and testing cycles designed to ensure correct functionality.
9. How Ground Truth for the Training Set Was Established
Given the information provided, the concept of a "ground truth for the training set" as it pertains to typical AI/ML development (e.g., labeled data for model training) is not applicable here. The ground truth for the development and verification of the software's core functions (counting, tagging) would have been established by the software's design specifications and the expected, correct outputs for given inputs during testing. This is standard software engineering practice rather than a data-driven ground truth for a machine learning algorithm.
§ 880.2740 Surgical sponge scale.
(a)
Identification. A surgical sponge scale is a nonelectrically powered device used to weigh surgical sponges that have been used to absorb blood during surgery so that, by comparison with the known dry weight of the sponges, an estimate may be made of the blood lost by the patient during surgery.(b)
Classification. Class I (general controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter, subject to the limitations in § 880.9. The device also is exempt from the current good manufacturing practice requirements of the quality system regulation in part 820 of this chapter, with the exception of § 820.180, with respect to general requirements concerning records, and § 820.198, with respect to complaint files.