(84 days)
The uPMR 790 system combines magnetic resonance diagnostic devices (MRDD) and Position Tomography (PET) scanners that provide registration of high resolution physiologic and anatomic information, acquired simultaneously and iso-centrically. The combined system maintains independent functionality of the MR and PET devices, allowing for single modality MR and/or PET imaging. The MR is intended to produce sagittal, transverse, coronal, and oblique cross sectional images, and that display internal anatomical structure and/or function of the head, body and extremities. Contrast agents may be used depending on the region of interest of the scan. The PET provides distribution information of PET radiopharmaceuticals within the human body to assist healthcare providers in assessing the metabolic and physiological functions. The combined system utilizes the MR for radiation-free attenuation correction maps for PET studies. The system provides inherent anatomical reference for the fused PET and MR images due to precisely aligned MR and PET image coordinate systems.
The uPMR 790 system is a combined Magnetic Resonance Diagnostic Device (MRDD) and Positron Emission Tomography (PET) scanner. It consists of components such as PET detector, 3.0T superconducting magnet, RF power amplifier, RF coils, gradient power amplifier, gradient coils, patient table, spectrometer, computer, equipment cabinets, power distribution system, internal communication system, vital signal module, and software etc. The uPMR 790 system provides simultaneous acquisition of high resolution metabolic and anatomic information from PET and MR. PET detectors are integrated into the MR bore for simultaneous, precisely aligned whole body MR and PET acquisition. The PET subsystem supports Time of Flight (ToF). The system software is used for patient management, data management, scan control, image reconstruction, and image archive. The uPMR 790 system is designed to conform to NEMA and DICOM standards.
The provided text describes modifications to the uPMR 790 system, focusing on new software features and updated specifications. The key study related to an AI module is the "WFI based head MRAC" (Water Fat Imaging based PET head attenuation correction).
Here's an breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
| Feature/Metric | Acceptance Criteria | Reported Device Performance and Notes |
|---|---|---|
| PET Imaging - Accuracy (AI-based MRAC) | Average SUVmean error (CTAC as reference) in major brain regions across the test patient cohort is within 10%. | Most brain regions having average error below 3%. (This fulfills the criteria, implying better performance than the threshold.) |
| PET Imaging - Sensitivity | 0cm: ≥14cps/kBq; 10cm: ≥14cps/kBq (for proposed device) | 0cm: ≥15cps/kBq; 10cm: ≥15cps/kBq (for predicate device) - Note 2 states that the proposed device sensitivity was updated to a better criterion due to an updated calibration factor, implying the proposed device also meets or exceeds the predicate's performance. |
| PET Imaging - NECR peak | ≥110kcps | ≥110kcps (Same as predicate) |
| PET Imaging - True peak | ≥300kcps | ≥360kcps (For predicate; Note 3 states the true peak value is updated to a wider criterion and will not affect system effectiveness, implying the proposed device meets this as well, and potentially with an even wider margin.) |
| PET Imaging - Scatter Fraction | ≤0.46 | ≤0.46 (Same as predicate) |
| PET Imaging - Image Quality (Accuracy) | Maximum value of the bias at or below NECR peak activity value: ≤10% | Maximum value of the bias at or below NECR peak activity value: ≤12% (for predicate). Note 4 states the accuracy specification for the proposed device updates to a better criterion due to physical correction optimization, implying it is ≤10%. |
| PET Imaging - Image Quality (Contrast Recovery coefficient) | 10mm: ≥45.0%; 13mm: ≥55.0%; 17mm: ≥55.0%; 22mm: ≥65.0%; 28mm: ≥65.0%; 37mm: ≥70.0% | Same as predicate. |
| PET Imaging - Image Quality (Noise) | 10mm: ≤9.0%; 13mm: ≤8.0%; 17mm: ≤7.0%; 22mm: ≤7.0%; 28mm: ≤7.0%; 37mm: ≤7.0% | Same as predicate. |
| PET Imaging - Image Quality (Relative lung error) | ≤10% | ≤16% (for predicate). Note 5 states the relative lung error specification for the proposed device updates to a better criterion due to verification, implying it is ≤10%. |
| PET Imaging - Time of Flight resolution | ≤560ps (This value is added, not a comparison to predicate, which listed N.A.) | The text states "Time of Fly resolution improve the image signal noise ratio," implying the device meets this new specification. |
| Safety (Surface Heating) | Consistent with NEMA MS 14-2019 (worst-case normal operating conditions for RF coil heating) | "The results for the surface heating test showed that proposed devices perform as well as or better than predicate devices." |
2. Sample Size Used for the Test Set and Data Provenance
- Test set for WFI-based head MRAC (AI module): 27 subjects (17 male, 10 female; Age: 15-78).
- Provenance: All subjects were Chinese.
- Data Independence: Test data from Center 2 (n=12) was initially excluded from the training set and from completely different subjects. Data from Center 1 (n=10) was collected almost 2 years after the training data's imaging date, also with different subjects. This confirms the test data was completely independent from the training data (prospective data collection from geographically and temporally distinct sources relative to the training set). The document implies a retrospective collection of these test cases from these two centers for the purpose of this validation.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications
- This information is not explicitly stated in the provided text for the WFI-based head MRAC study. The ground truth method is described as "three-compartment segmentation from CT images of the same person," and "image intensity threshold" for further segmentation, implying an objective, image-based ground truth rather than expert consensus on a subjective scale.
4. Adjudication Method for the Test Set
- Adjudication method is not applicable and not mentioned, as the ground truth for the AI module's performance was established via CT segmentation and image intensity thresholds rather than human readers.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done
- No, a multi-reader multi-case (MRMC) comparative effectiveness study was not conducted or reported in the provided text. The study focuses on the standalone performance of the AI module for attenuation correction, compared against CT-based ground truth (CTAC). There is no mention of human readers assisting with the AI or a comparison of human readers with and without AI assistance.
6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done
- Yes, a standalone performance evaluation was done for the "WFI based head MRAC" AI module. The performance was measured by comparing the SUVmean error of the AI-generated attenuation correction maps against CTAC (CT-based attenuation correction), which is considered the reference standard.
7. The Type of Ground Truth Used
- For the WFI-based head MRAC AI module, the ground truth used was "three-compartment segmentation from CT images of the same person" and subsequent separation using "image intensity threshold." This can be classified as a radiological/imaging-based ground truth derived from an existing gold standard imaging modality (CT).
8. The Sample Size for the Training Set
- Training dataset: Not explicitly stated as a single number. The text provides demographic information: Gender (76 male, 54 female), Age (17-83), Ethnicity (Chinese). This implies a total of 130 subjects in the training set (76 + 54).
9. How the Ground Truth for the Training Set Was Established
- The ground truth for the training set for the WFI-based head MRAC AI module was established using "pairs of WFI images and three-compartment segmentation from CT images of the same person." This means that for each subject in the training set, both WFI MR images and corresponding CT images were acquired, and the CT images were segmented into three compartments (air, cortical bone, mixed compartment) to serve as the ground truth for attenuation correction mapping. The AI module (Convolution Neural Network) was then trained to generate these segmentation masks from the WFI images.
§ 892.1200 Emission computed tomography system.
(a)
Identification. An emission computed tomography system is a device intended to detect the location and distribution of gamma ray- and positron-emitting radionuclides in the body and produce cross-sectional images through computer reconstruction of the data. This generic type of device may include signal analysis and display equipment, patient and equipment supports, radionuclide anatomical markers, component parts, and accessories.(b)
Classification. Class II.