(199 days)
The IMR reconstruction feature is intended as an alternative to standard reconstruction methods (filtered back projection) for the reconstruction of CT scanner data to produce diagnostic images. The IMR reconstruction feature is designed to reduce image noise, increase high-contrast spatial resolution, and improve low contrast detectability. IMR is designed to reduce dose required for diagnostic CT imaging, Image quality improvements and dose reduction depend on the clinical task, patient size, anatomical location, and clinical practice. IMR images will be used by a trained medical professional for diagnosis of clinical conditions in patients, including pediatrics and adults, who have been prescribed a CT scan as part of their clinical care
The IMR reconstruction feature is intended as an alternative to standard reconstruction methods (filtered back projection) for the reconstruction of CT scanner data to produce diagnostic images. The IMR reconstruction feature is designed to reduce image noise, increase high-contrast spatial resolution, and improve low contrast detectability. IMR is designed to reduce dose required for diagnostic CT imaging. Image quality improvements and dose reduction depend on the clinical task, patient size, anatomical location, and clinical practice. The IMR Software Application will reside on any Philips CT System that meets minimum software platform and hardware requirements. IMR enables the user to apply iterative reconstruction techniques to reconstruct raw CT data to generate diagnostic CT images. The use of IMR to reconstruct images may be done prospectively or retrospectively.
Here's a breakdown of the acceptance criteria and study information for the Philips IMR Software Application, based on the provided text:
1. Acceptance Criteria and Device Performance
| IMR Image Quality Parameter | Reported Device Performance (Lower Limit) | Reported Device Performance (Upper Limit) |
|---|---|---|
| Dose Reduction (with simultaneous improved image quality) | 60% lower radiation dose with 43% improvement in low contrast detectability and 83% less image noise. | 80% lower radiation dose with 80% improvement in low contrast detectability and 70% less image noise. |
| High-Contrast Spatial Resolution Improvement | 1.2x improvement | 1.7x improvement |
| Low-Contrast Detectability Improvement | 2.5x improvement | 3.6x improvement |
| Noise Reduction | Up to 90% image noise reduction | N/A (single value given) |
2. Sample Size and Data Provenance
- Test Set (Clinical Image Evaluation): 110 clinical image raw data sets.
- Data Provenance: Not explicitly stated (e.g., country of origin). The text refers to "clinical image raw data sets," implying retrospective data.
3. Number of Experts and Qualifications for Ground Truth (Clinical Image Evaluation)
- Number of Experts: A panel of eight physicians.
- Qualifications: Not explicitly stated beyond "physicians."
4. Adjudication Method (Clinical Image Evaluation)
- The text states: "Resulting data confirmed that the IMR software application provides diagnostic quality images and in majority of the cases, the physicians preferred the IMR images to standard (filter back projection) reconstruction." This suggests a comparative evaluation by the panel, likely involving subjective preference between FBP and IMR images, but a formal adjudication method like "2+1" or "3+1" is not specified.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- Was it done?: Yes, a "human observer study" was conducted for low contrast detectability (LCD), and a clinical image evaluation involving "a panel of eight physicians."
- Effect Size of human readers' improvement with AI vs. without AI assistance:
- Low Contrast Detectability (4-AFC Detection human observer study): The resulting data "confirmed the improved LCD using the IMR application," with an improvement range of 43% - 80% relative to filtered backprojection, as demonstrated through phantom-based tests.
- Clinical Image Evaluation: Physicians "preferred the IMR images to standard (filter back projection) reconstruction in the majority of the cases." No specific quantitative improvement (effect size) for human readers is given for this study beyond preference.
6. Standalone Performance (Algorithm Only)
- Was it done?: Yes, objective image quality testing was conducted using phantoms to determine noise, CT number uniformity, and high-contrast spatial resolution. The CT scan data was reconstructed by the IMR software application and compared to filtered backprojection. These are standalone performance metrics of the algorithm.
7. Type of Ground Truth Used
- Phantom-based tests: For noise, CT number uniformity, high-contrast spatial resolution, and low contrast detectability, phantom-based measurements were used as the ground truth, following methodologies like IEC 61223-3-5 and a 4-AFC detection human observer study with a low-contrast pin phantom.
- Clinical Image Evaluation: For the clinical image evaluation, the ground truth was based on the diagnosed quality and preferences of a panel of eight physicians, comparing IMR to standard (FBP) reconstructions. This is a form of expert consensus/subjective evaluation.
8. Sample Size for the Training Set
- The document does not provide information regarding the sample size used for training the IMR software application. The studies described are primarily for performance testing and validation.
9. How Ground Truth for the Training Set Was Established
- The document does not provide information on how ground truth was established for a training set, as it focuses on performance testing rather than development or training methodologies.
§ 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.