(185 days)
Philips Magnetic Resonance (MR) systems are Medical Electrical Systems indicated for use as a diagnostic device. This MR system enables trained physicians to obtain cross-sectional images and/or spectra of the internal structure of the head, body or extremities, in any orientation, representing the spatial distribution of protons or other nuclei with spin. Image appearance is determined by many different physical properties of the tissue and the anatomy, the MR scan technique applied, and presence of contrast agents for diagnostic imaging applications should be performed consistent with the approved labeling for the contrast agent. The trained clinical user can adjust the MR scan parameters to customize image appearance, accelerate image acquisition, and synchronize with the patient's breathing or cardiac cycle. The systems can use combinations of images to produce physical parameters, and related derived images, spectra, and measurements of physical parameters, when interpreted by a trained physician, provide information that may assist diagnosis and therapy planning. The accuracy of determined physical parameters depends on system and scan parameters, and must be controlled and validated by the clinical user. In addition the Philips MR systems provide imaging capabilities, such as MR fluoroscopy, to guide and evaluate interventional and minimally invasive procedures in the head, body and extremities. MR Interventional procedures, performed inside or adjacent to the Philips MR system, must be performed with MR Conditional or MR Safe instrumentation as selected and evaluated by the clinical user for use with the system configuration in the hospital. The appropriateness and use of information from a Philips MR system for a specific interventional procedure and specific MR system configuration must be validated by the clinical user.
The proposed Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems R11.0 are 60 cm and 70 cm bore 1.5 and 3.0 Tesla (1.5T and 3.0T) Magnetic Resonance Diagnostic Devices, hereafter to be known as Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems. This bundled abbreviated 510(k) submission will include modifications of the Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems as compared to our legally marketed devices Achieva, Intera, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems R5.7 (K193215, 04/10/2020). In this 510(k) submission, Philips Medical Systems Nederland B.V. will be addressing the following minor software enhancements to the proposed Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems when compared to the legally marketed predicate Achieva, Intera, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems R5.7 (K193215, 04/10/2020): 1. SmartSpeed AI 2. SmartSpeed MotionFree 3. SmartSpeed 3D FreeBreathing 4. SmartSpeed Implant 5. SmartSpeed DWI 6. MR Workspace 7. ISP MR Packages 8. Extended functionality Options This 510(k) submission will also address the following minor hardware enhancements: 1. Introduction of a graphical processing unit in the host recon computer for image reconstruction 2. Additional monitor as part of the operating console The supporting documentation provided for the proposed Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems, includes software and hardware modifications that are addressed in test reports for system level development project, Voyager. The proposed Achieva, Ingenia, Ingenia CX, Ingenia Elition and Ingenia Ambition MR Systems are intended to be marketed with the following pulse sequences and coils that are previously cleared by FDA: 1. mDIXON (K102344) 2. SWIp (K131241) 3. mDIXON-Quant (K133526) 4. MRE (K140666) 5. mDIXON XD (K143128) 6. O-MAR (K143253) 7. 3D APT (K172920) 8. Compatible System Coils
The provided text describes modifications to Philips MR systems, specifically the integration of "SmartSpeed AI" which combines previously cleared Compressed-SENSE with machine learning for improved image acquisition. The document focuses on demonstrating substantial equivalence to a predicate device rather than outright proving a novel device's performance against clinical endpoints.
Here's an analysis of the acceptance criteria and study data based on the provided text:
1. A table of acceptance criteria and the reported device performance:
The document doesn't explicitly present a formal "acceptance criteria table" with numerical targets. Instead, it describes the performance goals qualitatively, primarily focusing on "equivalent or better image quality" compared to images acquired without SmartSpeed AI but with longer scan times. The acceptance criteria essentially revolve around demonstrating that the new SmartSpeed AI feature does not negatively impact image quality and, ideally, improves it, especially at higher acceleration factors and lower SNR.
| Acceptance Criteria (Qualitative) | Reported Device Performance |
|---|---|
| Comparable or better results than data reconstructed without SmartSpeed AI ("fully sampled ground truth data" and "data reconstructed without SmartSpeed AI" are used as benchmarks). | "SmartSpeed AI does provide comparable or better results than the data reconstructed without SmartSpeed AI." |
| "SmartSpeed AI showed better alignment with the ground truth data for high acceleration factors and low SNR levels compared to the data reconstructed without SmartSpeed AI." | |
| Does not negatively impact image quality measures when acquired with reduced scan time. | "In vivo images were analyzed to confirm that SmartSpeed AI does not negatively impact image quality measures when acquired with reduced scan time." |
| Images with equivalent or better image quality when comparing SmartSpeed AI images acquired with shorter scan times to images without SmartSpeed AI acquired with longer scan times. | "The combined results of the comparison described above confirmed that the SmartSpeed AI feature provides images with equivalent or better image quality." |
2. Sample size used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective):
- Test Set Sample Size: The document does not specify a numerical sample size for the test set (number of images or patients). It mentions "a variety of datasets from different anatomies and image contrasts, varying SNR levels and acceleration factors" for pixel-wise comparison, and "in vivo images" for image quality analysis. For the reader evaluation study, it mentions "SmartSpeed AI images acquired across a variety of pulse sequences and anatomies."
- Data Provenance: Not explicitly stated. The phrase "in vivo images" and "a variety of datasets" implies real patient data, but the origin (e.g., country) is not mentioned, nor is whether the data was retrospective or prospectively collected for this study. Given it's a 510(k) for an upgrade, retrospective data is plausible.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g., radiologist with 10 years of experience):
- Number of Experts: "A reader evaluation study with US board certified radiologists was performed." The exact number of radiologists is not specified, only that it was plural ("radiologists").
- Qualifications of Experts: "US board certified radiologists." No further details on their experience level (e.g., years of experience, subspecialty) are provided.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set:
The document states: "Radiologists were asked to perform comparisons of SmartSpeed AI images acquired with shorter scan times and images without SmartSpeed AI acquired with longer scan times." It does not describe any specific adjudication method (e.g., consensus reading, majority vote) if there were multiple readers. It simply states "The combined results of the comparison described above confirmed..." suggesting an aggregation of individual reader opinions.
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:
- MRMC Study: A reader evaluation study was performed, which is a type of MRMC study. However, this study was not designed to measure "how much human readers improve with AI vs. without AI assistance." Instead, it was designed to compare the image quality of SmartSpeed AI images (shorter scan time) against non-SmartSpeed AI images (longer scan time), with human readers providing the comparison. The goal was to show non-inferiority or superiority in image quality, not an improvement in diagnostic performance of the human reader.
- Effect Size: No effect size regarding human reader improvement is reported because that was not the objective of the study. The study aimed to assess equivalent or better image quality of the AI-processed images.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
Yes, a standalone performance assessment was done. The document states:
- "A pixel-wise comparison was performed to confirm that SmartSpeed AI does provide comparable or better results than the data reconstructed without SmartSpeed AI."
- "SmartSpeed AI showed better alignment with the ground truth data for high acceleration factors and low SNR levels compared to the data reconstructed without SmartSpeed AI."
- "In vivo images were analyzed to confirm that SmartSpeed AI does not negatively impact image quality measures when acquired with reduced scan time."
These directly assess the algorithm's output (image quality) without human interpretation in the loop as the primary endpoint for these specific analyses.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
The ground truth for the pixel-wise comparison was "fully sampled ground truth data." This implies a reference image acquired with conventional, unaccelerated scanning techniques, which is considered the "true" or ideal image without any AI reconstruction. For the in vivo image quality assessment and reader study, the ground truth was essentially the "image quality" as perceived by "US board certified radiologists" in comparison to the non-AI enhanced, longer-scan-time images. It's a comparative ground truth based on expert perception rather than a definitive clinical diagnosis or pathology.
8. The sample size for the training set:
The document does not specify the sample size for the training set used for the "SmartSpeed AI" machine learning component.
9. How the ground truth for the training set was established:
The document mentions that SmartSpeed AI "combining the previously cleared and legally marketed feature Compressed-SENSE... with machine learning." Given the context of image reconstruction and enhancement, it's highly probable the training ground truth involved pairs of unaccelerated (or conventionally accelerated) MR images and corresponding undersampled or noisy MR data, allowing the AI to learn to reconstruct high-quality images from suboptimal inputs. However, the exact method for establishing this ground truth (e.g., specific image acquisition protocols, expert annotation for quality metrics) is not detailed in the provided text.
§ 892.1000 Magnetic resonance diagnostic device.
(a)
Identification. A magnetic resonance diagnostic device is intended for general diagnostic use to present images which reflect the spatial distribution and/or magnetic resonance spectra which reflect frequency and distribution of nuclei exhibiting nuclear magnetic resonance. Other physical parameters derived from the images and/or spectra may also be produced. The device includes hydrogen-1 (proton) imaging, sodium-23 imaging, hydrogen-1 spectroscopy, phosphorus-31 spectroscopy, and chemical shift imaging (preserving simultaneous frequency and spatial information).(b)
Classification. Class II (special controls). A magnetic resonance imaging disposable kit intended for use with a magnetic resonance diagnostic device only is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 892.9.