(262 days)
Zia™ Image Enhancement System is an image processing software that can be used for reducing noise in CT images. Enhanced images will be uploaded back to host/PACS systems and exist in conjunction to the original images. Zia™ is not intended for mammography applications. The device processing is not effective for lesion, mass, or abnormalities of sizes less than 2mm.
Zia™ Image Enhancement System is an image processing software that can be used for reducing noise in CT images. Zia™ image enhancement software is based on a core noise reduction algorithm that reduces noise in flat regions via a regularization process while keeping the edges via data fidelity constrains. The software, which is installed on a remote computer, receives DICOM images from CT host computer (Zia DICOM node needs to be configured on the scanner), automatically processes the received images and uploads the post processed images back on to the host computer and/or other PACS systems. Enhanced images exist in conjunction to the original images.
Here's a breakdown of the acceptance criteria and study details for the Zia™ Image Enhancement System based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance:
| Performance Metric | Acceptance Criteria | Reported Device Performance | Test Result |
|---|---|---|---|
| Noise Reduction | Reduces noise in processed images by at least 10% | Reduced noise in processed images by at least 10% | PASS |
| CT# (Signal) Accuracy | Keeps CT# (signal) accuracy within +/- 1.0 HU | Kept CT# (signal) accuracy within +/- 1.0 HU | PASS |
| High Contrast Resolution | Maintains (preserves) high contrast resolution | Maintained (preserved) high contrast resolution | PASS |
| Low Contrast Resolution | Maintains (preserves) low contrast resolution | Maintained (preserved) low contrast resolution | PASS |
2. Sample Size Used for the Test Set and Data Provenance:
- Sample Size: A total of 81 datasets were processed and analyzed.
- Data Provenance: The data was generated using an ACR CT PHANTOM (Model 464) on three different CT scanners: GE BrightSpeed 4-Slice, Siemens Sensation 16-Slice, and Philips Brilliance 64-Slice. The images were acquired following specific protocols (Head 120KV, Head 80KV, and Body 120KV) with varying mAs (150-350mAs) and slice thicknesses (1.25-5mm). This indicates a controlled, simulated environment using a phantom, not retrospective or prospective patient data from a specific country of origin.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications:
The document does not mention the use of human experts to establish ground truth for the test set. The ground truth appears to be based on physical measurements of the ACR CT phantom.
4. Adjudication Method for the Test Set:
Not applicable, as human experts were not used for establishing ground truth or evaluating the test set. The performance was measured quantitatively using the phantom.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
No, a multi-reader multi-case (MRMC) comparative effectiveness study was not performed. The study focuses on the technical performance of the image processing software itself, not its impact on human reader performance.
6. Standalone Performance Study:
Yes, a standalone study was performed. The device's performance (noise reduction, CT# accuracy, contrast resolution) was evaluated directly by analyzing the processed images from the CT phantom, without human-in-the-loop.
7. Type of Ground Truth Used:
The ground truth used was based on the physical characteristics and known properties of the ACR CT PHANTOM (Model 464). Measurements of CT# and noise were obtained from specific regions of interest (ROIs) within the phantom, and resolution was assessed based on the phantom's design.
8. Sample Size for the Training Set:
The document does not specify a separate training set or its sample size. The description of the device's core algorithm as reducing noise in flat regions while keeping edges suggests it's a rule-based or model-based algorithm, rather than a machine learning algorithm requiring a separate, large training set with annotated data.
9. How the Ground Truth for the Training Set Was Established:
Not explicitly stated. Given the description of the algorithm, it likely relies on mathematical principles and image processing techniques. If there was any "training" in a general sense, it would have involved developing and refining these algorithms based on general image characteristics rather than a labeled training dataset with a specific ground truth.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).