Search Results
Found 1 results
510(k) Data Aggregation
(145 days)
AI Segmentation uses CT images to segment patient anatomy for use in radiation therapy treatment planning. AI Segmentation utilizes a pre-defined set of organ structures in the following regions: head and neck, thorax, pelvis, abdomen. Segmentation results are subject to review and editing by qualified, expert radiation therapy treatment planners. Results of AI Segmentation are utilized in the Eclipse Treatment Planning System where it is the responsibility of a qualified physician to further review, edit as needed, and approve each structure.
AI segmentation is a web-based application, running in the cloud, that provides a combined deep learning and classical-based approach for automated segmentation of organs at risk, along with tools for structure visualization. This software medical device product is used by trained medical professionals and consists of a web application user interface where the results from the automated segmentation can be reviewed and selected for export into the compatible treatment planning system. AI Segmentation is not intended to provide clinical decisions, medical advice, or evaluations of radiation plans or treatment procedures.
The provided text describes that the AI Segmentation device does not include clinical data in its premarket submission. The document explicitly states: "No animal studies or clinical tests have been included in this pre-market submission." Therefore, it is not possible to provide acceptance criteria or a study proving the device meets the acceptance criteria using the requested information (e.g., sample size for the test set, number of experts, adjudication method, MRMC study, standalone performance, type of ground truth for test and training sets, and training set size).
Instead, the submission for AI Segmentation focused on non-clinical data, specifically software verification and validation testing, and conformance to relevant standards.
Here is what can be extracted from the document regarding the non-clinical evaluation:
1. Table of Acceptance Criteria and Reported Device Performance:
Since no clinical study data is available, a table of acceptance criteria and reported device performance in terms of clinical accuracy (e.g., Dice score, sensitivity, specificity) cannot be provided. The performance data presented is focused on software quality and adherence to regulatory standards.
| Acceptance Criterion (Non-Clinical) | Reported Device Performance |
|---|---|
| Conformance to applicable software requirements and specifications | "Test results demonstrate conformance to applicable requirements and specifications." (Page 5) |
| Software level of concern assessment | Assessed as "major" level of concern. (Page 5) |
| Conformance to IEC 62304 Edition 1.1 2015-06 (Medical device software - Software life cycle processes) | Conforms in whole or in part. (Page 5) |
| Conformance to IEC 62366-1 Edition 1.0 2015-02 (Application of usability engineering to medical devices) | Conforms in whole or in part. (Page 6) |
| Conformance to IEC 62083 Edition 2.0 2009-09 (Requirements for the safety of radiotherapy treatment planning systems) | Conforms in whole or in part. (Page 6) |
| Conformance to IEC 82304-1 Edition 1.0 2016-10 (Health software Part 1: General requirements for product safety) | Conforms in whole or in part. (Page 6) |
| Absence of Safety or Customer Intolerable Discrepancy Reports (DRs) | "There were no remaining discrepancy reports (DRs) which could be classified as Safety or Customer Intolerable." (Page 6) |
2. Sample size used for the test set and the data provenance: Not applicable, as no clinical test set data from patients was submitted. The evaluation was based on software testing.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts: Not applicable, as no clinical test set requiring expert ground truth was submitted.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set: Not applicable, as no clinical test set requiring adjudication was submitted.
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: No MRMC study was done, as explicitly stated that no clinical tests were included.
6. If a standalone (i.e. algorithm only without human-in-the loop performance) was done: The document states, "Segmentation results are subject to review and editing by qualified, expert radiation therapy treatment planners. Results of AI Segmentation are utilized in the Eclipse Treatment Planning System where it is the responsibility of a qualified physician to further review, edit as needed, and approve each structure." This indicates the device is intended for human-in-the-loop use. However, no clinical performance data (standalone or otherwise) was provided. The software verification and validation would have tested the algorithm's output without human intervention, but these were functional tests, not clinical performance studies.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.): Not applicable for clinical ground truth, as no clinical studies were submitted. For software verification, ground truth would be against predetermined functional requirements and expected outputs established during software development.
8. The sample size for the training set: Not provided in the document.
9. How the ground truth for the training set was established: Not provided in the document.
Ask a specific question about this device
Page 1 of 1