FDA DEVICE FACTS
Artificial Intelligence Renal Assessment (AIRA) Version 1
Aramis Global AI Consultancies, LLC
FDA-sourced product information for medical-device development research. Use the official FDA sources at the end of this page when the original public document wording is needed.医療機器開発の調査用に、FDA の公開情報から取った製品情報を示しています。公開文書の原文が必要なときは、このページの最後にある FDA の公式な出典をご覧ください。
Indications for use
Non-invasive processing of CT kidney images using automatic image segmentation and measurement of anatomical structures utilizing artificial intelligence/machine-learning algorithms. The system automatically calculates the renal parenchymal volume. Device Name Manufacturer 510(k) # Decision Date Ceevra Reveal 3 Ceevra Inc. K222676
04/25/2023 AIRA is intended for adult patients who require volumetric measurement of the kidneys for further clinical assessment.
Summary of Indications for Use AIRA and the predicate device have the same indication for use. Both devices are intended for non-invasive processing of images using automatic image segmentation and measurement of anatomical structures utilizing artificial intelligence/machine-learning algorithms. AIRA is intended for adult patients who require volumetric measurement of the kidneys for further clinical assessment. Summary of Technological Characteristic AIRA and the predicate device have similar technological characteristics. Both utilize artificial intelligence/machine-learning algorithms to perform automatic image segmentation and measurement of anatomical structures. AIRA is different from the predicate in that it uses images from CT and MRI while the subject device utilizes CT images only. General Safety and Effectiveness Concerns
The device labeling contains instructions for use and any necessary cautions and warnings to provide for safe and effective use of this device. Risk management is ensured via a risk analysis which is used to identify and mitigate potential hazards. Any potential hazards are controlled via software development, verification and validation testing. In addition, general controls of the FD&C Act, and special controls established for medical image management and processing system per 21 CFR 892.2050 are in place to further mitigate any safety and or effectiveness risks.
Assessment of Non-Clinical Performance Data
The AIRA Standalone Study was a retrospective study conducted to assess the performance of AIRA. A total of 119 computed tomography (CT) cases were retrospectively collected for use in this standalone study under the approved Aramis Protocol 66017 Data Collection Protocol. The cases were collected from 3 US sites at distinct geographic locations.
The study analyses included 113 cases. The sample was enriched to have at least 25% of cases with atrophied kidneys. Ground truth volume and segmentation were established by 2 radiologists; no adjudication was required. AIRA: Version 1.0 was used to process all cases and provide segmentation and RPV for each kidney. Accuracy of AIRA was measured by the Dice similarity coefficient (DSC) and the Hausdorff distance between AIRA and ground truth segmentation.
The mean Dice Similarity Coefficient (DSC) between the AIRA mask and the larger Ground Truther mask was 0.9092 for the Right kidney (95% CI: 0.9032, 0.9148) and 0.9114 for the Left kidney (95% CI: 0.9035, 0.9179). The mean Hausdorff Distance (HD) between the AIRA mask and the larger Ground Truther mask was 9.9743 mm for the Right kidney (95% CI: 8.4073, 12.5967) and 9.9395 mm for the Left kidney (95% CI: 7.8926, 13.6073). Mean 95th Percentile HDs (HD95s) were 3.7734 mm (95% CI: 3.5544, 4.0104) and 3.9151 mm (95% CI: 3.4522, 4.6661) for the Right and Left kidneys, respectively.
Device description, AI/software, PCCP, evaluation and limitations are available with an account: seven days of full access after free registration, then with the paid plan.機器の説明、AI・ソフトウェア、PCCP、評価、制限事項は、アカウントでご覧いただけます。無料登録から 7 日間はすべて利用でき、その後は有料プランでご利用いただけます。
OFFICIAL FDA SOURCES
FDA Record and Public Document Status
Record and document link states are stored separately.FDA Record と公開文書のリンクの状態は、別々に記録しています。