(161 days)
QSM software (QSMetric™) is intended for use in the post-acquisition image enhancement of 3D MR images of the brain acquired using a gradient-echo sequence at 1.5T, 3T and 7T field strengths. QSM uses phase information to enhance contrast between tissues presenting magnetic susceptibility differences, such as deoxygenated blood, iron or calcium deposits. When used in combination with other clinical information, QSM may aid the qualified physicians in visualizing tissue structures with magnetic susceptibility contrasts and measuring their susceptibility values.
The product QSMetric™, also referred to as QSM software, postprocesses gradient echo magnetic resonance (MR) images to depict tissue magnetic susceptibility contrast (local difference). Tissue susceptibility contrast sources include highly paramagnetic iron presented in deoxyhemoglobin, ferritin and hemosiderin, and diamagnetic calcification. Susceptibility contrast material of tissue in the MR scanner generates its own magnetic field according to the convolution law in magnetism. This tissue field with its dispersion in space causes MR image signal magnitude loss, creating contrasts in magnitude images. Therefore, the magnitude image is commonly used to indicate the presence of nearby tissue susceptibility contrast.
In addition to magnitude images, a gradient echo MR scan results in also phase images. The tissue field causes MR image signal phase accrual, creating contrasts in phase images. The phase image of gradient echo MR data is the product of echo time and tissue field. Accordingly, the phase images from a gradient echo MR scan is processed using QSM to enhance the depiction of tissue susceptibility contrasts.
QSM software works in conjunction with any FDA cleared third-party DICOM viewer as an image postprocessing solution in radiological service.
The provided text details the FDA 510(k) clearance for the QSM software (QSMetric™) but does not contain a specific "acceptance criteria table" or the full study details typically found in a clinical study report. The document focuses on demonstrating substantial equivalence to a predicate device (SWIp from Philips).
However, I can extract the information provided about the device's performance, the studies conducted, and how ground truth was established, to the best of what's available in the text.
Here's a breakdown of the requested information based on the provided FDA 510(k) summary:
Acceptance Criteria and Device Performance
As a specific table of "acceptance criteria" with quantitative metrics is not explicitly provided in the document for the device's performance itself (e.g., sensitivity, specificity, accuracy for a diagnostic task), the information below is inferred from the language used in the "SE-Nonclinical performance data" and "SE-Clinical performance data" sections. The acceptance criteria described are primarily related to engineering performance testing and clinical user needs testing to demonstrate substantial equivalence to the predicate, rather than direct diagnostic performance against a gold standard for a specific clinical condition.
| Acceptance Criterion (Inferred from text) | Reported Device Performance and Conclusion |
|---|---|
| Nonclinical Performance (Engineering Testing) | |
| - Consistent production of results according to intended use. | "All predefined acceptance criteria for the engineering performance testing were met for all test cases across different scanner manufacturers. The results from the nonclinical testing performed on the QSM software demonstrate that the QSM software produces results consistently according to its intended use..." |
| - Substantially equivalent to combining information from SWIp magnitude | "...and is substantially equivalent to combining information from SWIp both magnitude and phase images output from the predicate device." |
| and phase images (predicate). | |
| Clinical Performance (Clinical User Needs Testing) | |
| - Acceptable output image quality. | "All predefined acceptance criteria for clinical validation testing, including clinical user needs testing, as a part of the QSM performance validation testing efforts, were met across all test cases. The results of the clinical validation related testing performed on the QSM software demonstrate that output image quality are acceptable, all clinical user needs are met..." |
| - All clinical user needs met. | "...all clinical user needs are met..." |
| - Substantially equivalent to combining information from SWIp magnitude | "...and QSM is substantially equivalent to combining information from SWIp both magnitude and phase images output from the predicate device." The document explicitly states "The subject device of this premarket notification, QSM software, did not require clinical studies to support substantial equivalence to the predicate device," indicating that the "clinical validation testing" mentioned was not a traditional clinical trial assessing diagnostic accuracy, but rather verification of user needs and perceived image quality in a clinical context. |
| and phase images (predicate). |
Study Details
-
Sample sizes used for the test set and the data provenance:
- Test Set Sample Size: Not explicitly stated with a specific number of cases for either non-clinical or clinical validation. The text uses general terms like "all test cases" for both engineering performance testing and clinical validation testing.
- Data Provenance: Not specified (e.g., country of origin). The document indicates testing across "different scanner manufacturers" (General Electric, Philips, Siemens, and United Imaging), implying data from various MRI systems, but not their geographical origin. It does not explicitly state whether the data was retrospective or prospective.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- Not applicable/Not specified. The ground truth, as described, was primarily based on the expected output and performance of an image processing software in comparison to a predicate device, rather than a diagnostic 'ground truth' established by expert consensus for a specific disease or condition. The "clinical user needs testing" suggests involvement of users (presumably qualified physicians as per the intended use), but the number and specific qualifications for establishing 'ground truth' for this type of software validation are not detailed.
-
Adjudication method for the test set:
- Not applicable/Not specified. Given the nature of the software (image enhancement/post-processing for visualization and measurement of susceptibility values), typical diagnostic adjudication methods (like 2+1 or 3+1 for disease presence) are not described as part of the validation presented. The "acceptance criteria" were met through engineering and user needs testing, implying a different validation approach.
-
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. The document explicitly states: "The subject device of this premarket notification, QSM software, did not require clinical studies to support substantial equivalence to the predicate device." Therefore, an MRMC study comparing human readers with and without AI assistance was not conducted or presented.
-
If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
- Yes, implicitly. The "nonclinical performance data" and "engineering performance testing" describe the algorithm's performance in producing consistent results and being substantially equivalent to the predicate, which would involve standalone assessment of the software's output. The "clinical validation testing" focused on "output image quality" and "clinical user needs met," which also points to an assessment of the algorithm's output, though likely interpreted by human users.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
- The "ground truth" for this device's validation appears to be:
- For engineering performance: The expected computational output and consistency across different scanner types, and the comparison to the existing, cleared predicate device's output (SWIp magnitude and phase images).
- For "clinical validation" (user needs): Acceptable image quality and meeting predefined clinical user needs, likely assessed qualitatively or semi-quantitatively by qualified personnel, in comparison to the output of the predicate device. It's not a ground truth for a disease diagnosis but for the quality and utility of the enhanced images.
- The "ground truth" for this device's validation appears to be:
-
The sample size for the training set:
- Not specified. The document outlines verification and validation procedures but does not provide details on the training set or its size, which is typical for a 510(k) summary focused on substantial equivalence rather than detailed algorithm development.
-
How the ground truth for the training set was established:
- Not specified. As the training set size or details are not provided, the method for establishing its ground truth is also not mentioned.
§ 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.