(205 days)
ADMIRE is a CT reconstruction software. The end user can choose to apply either ADMIRE or the weighted filter backprojection (WFBP) to the acquired raw data. Depending on the clinical task, patient size, anatomical location, and clinical practice, the use of ADMIRE can help to reduce radiation dose while maintaining pixel noise, low contrast detectability and high contrast resolution. Phantom measurements showed that high contrast resolution and pixel noise are equivalent between full dose WFBP images and reduced dose ADMIRE images. Additionally, ADMIRE can reduce spiral artifacts by using iterations going back and forth between image space and raw data space.
Images reconstructed with ADMIRE are not intended to be evaluated with syngo Osteo CT or syngo Calcium Scoring.
Siemens ADMIRE is an extension of the previously cleared Sinogram Affirmed Iterative Reconstruction (SAFIRE) reconstruction algorithm. ADMIRE is a software option for CT operating systems that provides an improved image quality or reciprocally can allow the physician to acquire scans with reduced radiation dose without reduction of image quality compared to today's standard.
ADMIRE is designed to improved reconstructed image quality through the integration of additional processing steps in image reconstruction. These additional steps result in the following improvements in image quality:
- . Higher pixel noise reduction
- A noise texture closer to filtered back projection (FBP) .
- . Improved resolution for high contrast edges
Here's an analysis of the provided text regarding the acceptance criteria and the study that proves the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
The document doesn't explicitly state quantitative acceptance criteria for device performance. Instead, it frames the performance improvements of ADMIRE relative to its predicate device (SAFIRE) and traditional methods (FBP/WFBP). The "acceptance criteria" are implied to be the successful demonstration of these improvements and equivalence for certain metrics.
| Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|
| Higher pixel noise reduction (especially in thicker slices) | ADMIRE provides higher pixel noise reduction in thicker slices (e.g., 3mm and 5mm) compared to SAFIRE. Additionally, it helps maintain pixel noise at reduced radiation doses compared to full dose WFBP images. |
| Noise texture closer to filtered back projection (FBP) | ADMIRE results in a noise texture closer to FBP, with fewer outliers, compared to SAFIRE. |
| Improved resolution for high contrast edges | ADMIRE shows improved resolution for high contrast edges compared to SAFIRE and weighted filtered back projection (WFBP). Phantom measurements showed that high contrast resolution is equivalent between full dose WFBP images and reduced dose ADMIRE images. |
| Ability to reduce radiation dose while maintaining image quality | The use of ADMIRE can help to reduce radiation dose while maintaining pixel noise, low contrast detectability, and high contrast resolution. Phantom measurements demonstrated equivalence in high contrast resolution and pixel noise between full dose WFBP images and reduced dose ADMIRE images. |
| Reduction of spiral artifacts | ADMIRE can reduce spiral artifacts by using iterations going back and forth between image space and raw data space. |
| Conformance with safety and performance standards | ADMIRE fulfills requirements of various standards (e.g., ISO 14971, IEC 62304, IEC 60601-1-4, IEC 60601-1-6, NEMA DICOM PS 3.1-3.18). Risk analysis was completed, and risk control implemented. EMC/electrical safety evaluated according to IEC Standards. All software specifications met acceptance criteria. Identified risk of electrical hazards mitigated. |
| Substantial equivalence to predicate device (SAFIRE) | Siemens is of the opinion that ADMIRE does not introduce any new potential safety risk and is substantially equivalent to and performs as well as the predicate devices. |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: Not explicitly stated for specific tests. The document mentions "phantom testing" and "sample clinical images." No specific number of patients or imaging studies are provided.
- Data Provenance: The document does not specify the country of origin. It indicates that "phantom testing" was conducted and "sample clinical images were also provided within the submission." The general nature of the testing suggests it's likely internal Siemens data. The testing involves "simulated body and head phantoms."
- Retrospective/Prospective: The nature of the testing described (phantom testing, bench testing, verification/validation) suggests a controlled, likely retrospective analysis of specific data, or controlled prospective phantom acquisitions. There's no mention of a large-scale, prospective clinical trial with human subjects.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
This information is not provided in the document. The evaluation primarily relies on phantom measurements and technical performance metrics (noise reduction, resolution, noise texture) rather than a subjective human reader assessment against a clinical ground truth.
4. Adjudication Method for the Test Set
This information is not provided. Given the focus on objective phantom measurements and technical image quality metrics, a formal human reader adjudication method like 2+1 or 3+1 is unlikely to have been used in the primary performance evaluations described.
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, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study with human readers assessing AI-assisted vs. non-AI-assisted images is not mentioned or described in this 510(k) summary. The document focuses on the technical improvements of the reconstruction algorithm itself.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) Was Done
Yes, the performance described is primarily that of the standalone algorithm (ADMIRE). The improvements in pixel noise reduction, noise texture, high contrast resolution, and artifact reduction are inherent to the algorithm's processing of raw data. The statement "The end user can choose to apply either ADMIRE or the weighted filter back-projection (WFBP) to the acquired raw data" further reinforces its standalone nature as a reconstruction option.
7. The Type of Ground Truth Used
The ground truth used appears to be:
- Physical Phantom Measurements: For metrics like pixel noise, high contrast resolution, and low contrast detectability. This involves comparing the device's output against known physical properties or reference measurements from phantoms.
- Reference Reconstruction Methods: Comparisons are made against established methods like Filtered Back Projection (FBP) and Weighted Filtered Back Projection (WFBP) to demonstrate improvements or equivalence.
- Compliance with Standards: Verification against various international standards for medical devices and software (ISO, IEC, NEMA).
There is no mention of "expert consensus," "pathology," or "outcomes data" being used as ground truth for the performance evaluation described in this summary.
8. The Sample Size for the Training Set
This information is not provided. The document outlines changes to a reconstruction algorithm rather than a machine learning model that would typically have a distinct training set. While reconstruction algorithms are developed and refined using data, the document doesn't use the terminology of "training set" in the context of deep learning.
9. How the Ground Truth for the Training Set Was Established
This information is not provided, and the concept of a "training set" and associated ground truth, as typically understood in machine learning, is not discussed in this summary. The development process likely involved engineering and optimization against known physical models and existing reconstruction results, rather than labeled training data for a learning algorithm.
§ 892.1750 Computed tomography x-ray system.
(a)
Identification. A computed tomography x-ray system is a diagnostic x-ray system intended to produce cross-sectional images of the body by computer reconstruction of x-ray transmission data from the same axial plane taken at different angles. This generic type of device may include signal analysis and display equipment, patient and equipment supports, component parts, and accessories.(b)
Classification. Class II.