Search Results

DentiqAir is a software application for the visualization of imaging information of the oralmaxillofacial region. The imaging data originates from medical scanners such as CT or CBCT scanners. The dental professionals' planning data may be exported from DentiqAir and used as input data for CAD or Rapid Prototyping Systems.

DentiqAir is a pure software device applied for the visualization of imaging information of the ear-nose-throat (ENT) region and oral-maxillofacial region. The imaging data originates from medical scanners such as CT or Cone Beam - CT (CBCT) scanners. This information can be complemented by the imaging information from optical impression systems. The medical professionals' input information and planning data may be exported from Dentiq Air to be used for CAD or Rapid Prototyping Systems.

The provided text describes the 510(k) premarket notification for the DentiqAir device. While it mentions performance tests, it does not include a detailed table of acceptance criteria and reported device performance for all features, nor does it provide a full breakdown of the test set, expert involvement, or MRMC study results typically found in comprehensive performance studies for AI/ML-driven devices.

However, based on the non-clinical testing section, we can infer some information regarding the performance and acceptance criteria for specific functionalities.

Here's a breakdown of the requested information based on the provided text:

1. Table of Acceptance Criteria and the Reported Device Performance:

The document primarily focuses on accuracy tests for measurements made using the device against phantom data. It does not provide performance metrics for segmentation accuracy (e.g., Dice score, Jaccard index) which might be expected for segmentation features.

Note: The phrasing "Less than True Value and more than Dolphin Imaging average" for Volume is a bit ambiguous regarding exact numerical targets, but it implies a comparative target against a predicate device's expected performance. The document states that the test results support substantial equivalence, implying these criteria were met.

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

• Sample Size: The document states that accuracy tests were conducted "from loaded CT datasets using phantom." It does not specify the number of CT datasets or phantoms used for this testing.
• Data Provenance: The data used was from "phantom" studies, meaning simulated or controlled anatomical models, not human patient data. The country of origin is not explicitly stated for the phantom data, but the submitter is from the Republic of Korea. It is retrospective in nature, as it's a pre-market submission based on completed testing.

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

• Number of Experts: Not applicable. The ground truth for the phantom accuracy tests was established by the known true values of the phantom itself, not by expert consensus.
• Qualifications of Experts: N/A as expert consensus was not the method for establishing ground truth for the stated performance tests.

4. Adjudication Method for the Test Set:

• Adjudication Method: Not applicable. Given the ground truth for the measurement accuracy tests was based on the known physical properties of the phantom, no human adjudication was necessary for these specific performance metrics.

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: No. The document explicitly states: "Clinical testing is not a requirement and has not been performed." The performance tests described are strictly non-clinical and focus on software functionality and measurement accuracy. This type of study would be highly relevant for devices intended to assist human readers in diagnosis or treatment planning.

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

• Standalone Performance: For the measurement accuracy tests (Length, Angle, HU, Volume), the described testing does reflect standalone performance as it compares the device's measurements directly to the phantom's true values, without human intervention in the measurement process itself.
• The software also includes "segmentation features," and the document states, "Performance testing has been used to validate the safety and effectiveness of the DentiqAir segmentation features in comparison to the predicate devices." However, no specific quantitative standalone performance metrics (e.g., Dice coefficient, precision, recall) for segmentation are provided in this summary.

7. The Type of Ground Truth Used:

• Ground Truth: For the measurement accuracy tests, the ground truth was based on the known physical properties (true values) of the phantom used for testing.

8. The Sample Size for the Training Set:

• Training Set Sample Size: Not provided. The document is a 510(k) summary for a software device. While it mentions "software verification and validation testing activities" including "code review, integration review, and dynamic tests," and "performance tests," it does not discuss the training or development of any AI/ML models within the software or the data used for such purposes. The device's segmentation algorithm is mentioned as "Water Shed (a type of graph-cut algorithm)," which is a traditional image processing algorithm rather than a deep learning model requiring a specific training set with labelled data in the sense of modern AI/ML.

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

• Ground Truth for Training Set: Not applicable. As the document does not describe the use of an AI/ML model that requires a labelled training set in the contemporary sense, the establishment of ground truth for a training set is not discussed. The Water Shed algorithm does not require labeled training data in the same way a deep learning model would.

Ask a specific question about this device