Search Results

The uMR 680 system is indicated for use as a magnetic resonance diagnostic device (MRDD) that produces sagittal, transverse, coronal, and oblique cross sectional images, and spectroscopic images, and that display internal anatomical structure and/or function of the head, body and extremities.

These images and the physical parameters derived from the images when interpreted by a trained physician yield information that may assist the diagnosis. Contrast agents may be used depending on the region of interest of the scan.

The uMR 680 is a 1.5T superconducting magnetic resonance diagnostic device with a 70cm size patient bore. It consists of components such as magnet, RF power amplifier, RF coils, gradient power amplifier, gradient coils, patient table, spectrometer, computer, equipment cabinets, power distribution system, internal communication system, and vital signal module etc. The uMR 680 Magnetic Resonance Diagnostic Device is designed to conform to NEMA and DICOM standards.

This traditional 510(k) is to request modifications for the cleared uMR 680(K240744). The modifications performed on the uMR 680 in this submission are due to the following changes that include:
(1) Addition of RF coils and corresponding accessories: Breast Coil -12, Biopsy Configuration, Head Coil-16, Positioning Couch-top, Coil Support.
(2) Deletion of VSM (Wireless UIH Gating Unit REF 453564324621, ECG module Ref 989803163121, SpO2 module Ref 989803163111).
(3) Modification of the dimensions of Detachable table: from width 826mm, height 880mm,2578mm to width 810mm, height 880mm, length 2505mm.
(4) Addition and modification of pulse sequences
a) New sequences: gre_snap, gre_quick_4dncemra, gre_pass, gre_mtp, gre_trass, epi_dwi_msh, epi_dti_msh, svs_hise.
b) Added associated options for certain sequences: fse(add Silicone-Only Imaging, MicroView, MTC, MultiBand), fse_arms(add Silicone-Only Imaging), fse_ssh(add Silicone-Only Imaging), fse_mx(add CEST, T1rho, MicroView, MTC), fse_arms_dwi(add MultiBand), asl_3d(add multi-PLD), gre(add T1rho, MTC, output phase image), gre_fsp(add FSP+), gre_bssfp(add CASS, TI Scout), gre_fsp_c(add 3D LGE, DB/GB PSIR), gre_bssfp_ucs(add real time cine), gre_fq(add 4D Flow), epi_dwi(add IVIM), epi_dti(add DKI, DSI).
c) Added additional accessory equipment required for certain sequences: gre_bssfp(add Virtual ECG Trigger).
d) Name change of certain sequences: gre_fine(old name: gre_bssfp_fi).
e) Added applicable body parts: gre_ute, gre_fine, fse_mx.
(5) Addition of imaging reconstruction methods: AI-assisted Compressed Sensing (ACS), Spark artifact Correction (SparkCo).
(6) Addition of imaging processing methods: Inline Cardiac Function, Inline ECV, Inline MRS, Inline MOCO, 4D Flow, SNAP, CEST, T1rho, FSP+, CASS, PASS, MTP.
(7) Addition of workflow features: TI Scout, EasyCrop, ImageGuard, Mocap, EasyFACT, Auto Bolus tracker, Breast Biopsy and uVision.
(8) Modification of workflow features: EasyScan(add applicable body parts)

The modification does not affect the intended use or alter the fundamental scientific technology of the device.

The provided FDA 510(k) clearance letter and summary for the uMR 680 Magnetic Resonance Imaging System outlines performance data for several new features and algorithms.

Here's an analysis of the acceptance criteria and the studies that prove the device meets them for the AI-assisted Compressed Sensing (ACS), SparkCo, Inline ED/ES Phases Recognition, and Inline MOCO algorithms.

1. Table of Acceptance Criteria and Reported Device Performance

• Age: 0.92-0.93
• Gender: 0.92
• Ethnicity: 0.91-0.92
• BMI: 0.91-0.95
• Magnetic field strength: 0.92-0.93
• Disease conditions: 0.91-0.93 |
  | | Dice Coefficient (Left Ventricular Myocardium after Motion Correction) Cardiac Dark Blood Images | The average Dice coefficient of the left ventricular myocardium after motion correction is greater than 0.87. | The average Dice coefficient of the left ventricular myocardium after motion correction is 0.96, which is greater than 0.87. Subgroup analysis also showed good generalization:
• Age: 0.95-0.96
• Gender: 0.96
• Ethnicity: 0.95-0.96
• BMI: 0.96-0.98
• Magnetic field strength: 0.96
• Disease conditions: 0.96-0.97 |

2. Sample Size Used for the Test Set and Data Provenance

• AI-assisted Compressed Sensing (ACS):
  • Sample Size: 1724 samples from 35 volunteers.
  • Data Provenance: Diverse demographic distributions (gender, age groups, ethnicity, BMI) covering various clinical sites and separated time periods. Implied to be prospective or a carefully curated retrospective set, collected specifically for validation on the uMR 680 system, and independent of training data.
• SparkCo:
  • Simulated Spark Testing Dataset: 159 spark slices (generated from spark-free raw data).
  • Real-world Spark Testing Dataset: 59 cases from 15 patients.
  • Data Provenance: Real-world data acquired from uMR 1.5T and uMR 3T scanners, covering representative clinical protocols. The report specifies "Asian" for 100% of the real-world dataset's ethnicity, noting that performance is "irrelevant with human ethnicity" due to the nature of spark signal detection. This is retrospective data.
• Inline ED/ES Phases Recognition:
  • Sample Size: 95 cases from 56 volunteers.
  • Data Provenance: Includes various ages, genders, field strengths (1.5T, 3.0T), disease conditions (NOR, MINF, DCM, HCM, ARV), and ethnicities (Asian, White, Black). The data is independent of the training data. Implied to be retrospective from UIH MRI systems.
• Inline MOCO:
  • Sample Size: 287 cases in total (105 cardiac perfusion images from 60 patients, 182 cardiac dark blood images from 33 patients).
  • Data Provenance: Acquired from 1.5T and 3T magnetic resonance imaging equipment from UIH. Covers various ages, genders, ethnicities (Asian, White, Black, Hispanic), BMI, field strengths (1.5T, 3.0T), and disease conditions (Positive, Negative, Unknown). The data is independent of the training data. Implied to be retrospective from UIH MRI systems.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts

• AI-assisted Compressed Sensing (ACS):
  • Number of Experts: More than one (plural "radiologists" used).
  • Qualifications: American Board of Radiologists certificated physicians.
• SparkCo:
  • Number of Experts: One expert for real-world SparkCo evaluation.
  • Qualifications: "one experienced evaluator." (Specific qualifications like board certification or years of experience are not provided for this specific evaluator).
• Inline ED/ES Phases Recognition:
  • Number of Experts: Not explicitly stated for ground truth establishment ("gold standard phase indices"). It implies a single, established method or perhaps a consensus by a team, but details are missing.
• Inline MOCO:
  • Number of Experts: Three licensed physicians.
  • Qualifications: U.S. credentials.

4. Adjudication Method for the Test Set

• AI-assisted Compressed Sensing (ACS): Not explicitly stated, but implies individual review by "radiologists" to rate diagnostic quality.
• SparkCo: For the real-world dataset, evaluation by "one experienced evaluator."
• Inline ED/ES Phases Recognition: Not explicitly stated; "gold standard phase indices" are referenced, implying a pre-defined or established method without detailing a multi-reader adjudication process.
• Inline MOCO: "Finally, all ground truth was evaluated by three licensed physicians with U.S. credentials." This suggests an adjudication or confirmation process, but the specific method (e.g., 2+1, consensus) is not detailed beyond "evaluated by."

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

• No MRMC comparative effectiveness study was explicitly described to evaluate human reader improvement with AI assistance. The described studies focus on the standalone performance of the algorithms or a qualitative assessment of images by radiologists for diagnostic quality.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

• Yes, standalone performance was done for all listed algorithms.
  • ACS: Evaluated quantitatively (SNR, Resolution, Contrast, Uniformity, Structure Measurement) and then qualitatively by radiologists. The quantitative metrics are standalone.
  • SparkCo: Quantitative metrics (Detection Accuracy, PSNR) and qualitative assessment by an experienced evaluator. The quantitative metrics are standalone.
  • Inline ED/ES Phases Recognition: Evaluated quantitatively as the error between algorithmic output and gold standard. This is a standalone performance metric.
  • Inline MOCO: Evaluated using the Dice coefficient, which is a standalone quantitative metric comparing algorithm output to ground truth.

7. The Type of Ground Truth Used

• AI-assisted Compressed Sensing (ACS):
  • Quantitative: Fully-sampled k-space data transformed to image space.
  • Clinical: Radiologist evaluation ("American Board of Radiologists certificated physicians").
• SparkCo:
  • Spark Detection Module: Location of spark points (ground truth for simulated data).
  • Spark Correction Module: Visual assessment by "one experienced evaluator."
• Inline ED/ES Phases Recognition: "Gold standard phase indices" (method for establishing this gold standard is not detailed, but implies expert-derived or a highly accurate reference).
• Inline MOCO: Left ventricular myocardium segmentation annotated by a "well-trained annotator" and "evaluated by three licensed physicians with U.S. credentials." This is an expert consensus/pathology-like ground truth.

8. The Sample Size for the Training Set

• AI-assisted Compressed Sensing (ACS): 1,262,912 samples (from a variety of anatomies, image contrasts, and acceleration factors).
• SparkCo: 24,866 spark slices (generated from 61 spark-free cases from 10 volunteers).
• Inline ED/ES Phases Recognition: Not explicitly provided, but stated to be "independent of the data used to test the algorithm."
• Inline MOCO: Not explicitly provided, but stated to be "independent of the data used to test the algorithm."

9. How the Ground Truth for the Training Set Was Established

• AI-assisted Compressed Sensing (ACS): Fully-sampled k-space data were collected and transformed to image space as the ground-truth. All data were manually quality controlled.
• SparkCo: "The training dataset for the AI module in SparkCo was generated by simulating spark artifacts from spark-free raw data... a total of 24,866 spark slices, along with the corresponding ground truth (i.e., the location of spark points), were generated for training." This indicates a hybrid approach using real spark-free data to simulate and generate the ground truth for spark locations.
• Inline ED/ES Phases Recognition: Not explicitly provided.
• Inline MOCO: Not explicitly provided.

Ask a specific question about this device

The uPMR 790 system combines magnetic resonance diagnostic devices (MRDD) and Positron Emission Tomography (PET) scanners that provide registration and fusion 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 spectroscopic images, and that display internal anatomical structure and/or function of the head, body and extremities. Contrast agents may be used depending on the reqion 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.

This traditional 510(k) is to request modifications for the cleared uPMR 790(K222540). The modifications performed on the uPMR 790 (K222540) in this submission are due to the following changes that include:

• (1) Addition of RF coils: SuperFlex Body 24, SuperFlex Large -12, SuperFlex Small -12.
• (2) Addition and modification of pulse sequences:
  • (a) New sequences: gre fine, fse arms dwi, fse dwi, fse mars sle, grase, gre_bssfp_ucs, gre_fq, gre_pass, gre_quick_4dncemra, gre_snap, gre_trass, gre_rufis, epi_dwi_msh, svs_wfs, svs_stme.
  • (b) Added Associated options for certain sequences: QScan, MultiBand, Silicon-Only Imaging, MoCap-Monitoring, T1rho, CEST, Inline T2 mapping, CASS, inline FACT, uCSR, FSP+, whole heart coronary angiography imaging, mPLD (Only output original control/labeling images and PDw(Proton Density weighted) images, no quantification images are output).
  • (c) Name change of certain sequences: gre ute(old name: gre ute sp), svs_press(old name: press),svs_steam(old name: steam), csi_press(old name: press), csi hise(old name: hise).
• (3) Addition of MR imaging processing methods: 2D Flow, 4D Flow, SNAP, CEST, T1rho, FSP+, CASS, PASS, Inline T2 Mapping and DeepRecon.
• (4) Addition and modification of PET imaging processing methods:
  • (a) The new PET imaging processing methods: Hyper DPR (also named HYPER AiR) and Digital Gating (also named Self Gating).
  • (b) The modified method: HYPER Iterative.
• (5) Addition of MR image reconstruction methods: AI-assisted Compressed Sensing (ACS).
• (6) Addition and modification of workflow features:
  • (a) The new workflow features: EasyCrop, MoCap-Monitoring and QGuard-Imaging.
  • (b) The modified workflow feature: EasyScan.
• (7) Addition Spectroscopy: Liver Spectroscopy, Breast Spectroscopy.
• (8) Additional function: MR conditional implant mode.

The provided text does not contain detailed acceptance criteria for the uPMR 790 device in the format of a table, nor does it describe a specific study proving the device meets these criteria in a comparative effectiveness study or standalone performance study as would typically be presented for an AI/ML medical device.

The document is a 510(k) summary, which focuses on demonstrating substantial equivalence to a predicate device rather than providing a detailed clinical study report with specific performance metrics against acceptance criteria.

However, based on the information available, I can extract and infer some aspects related to acceptance criteria and the performance study:

Inferred Acceptance Criteria and Reported Device Performance (based on provided text):

The device is an integrated MR-PET system. The modifications primarily involve new RF coils, pulse sequences, imaging processing methods, and workflow features. The performance data section describes non-clinical testing to verify that the proposed device met design specifications and is Substantially Equivalent (SE) to the predicate device.

While explicit quantitative acceptance criteria are not tabulated, the text implies that the performance of the modified device (uPMR 790) must be at least equivalent to, or better than, the predicate and reference devices regarding image quality and functionality.

Specifically for the new or modified features related to AI/ML (DeepRecon and ACS), the implicit acceptance criteria appear to be:

• DeepRecon:
  • Equivalence in performance to DeepRecon on the uMR Omega.
  • Better performance than NADR (No DeepRecon) in SNR and resolution.
  • Maintenance of image qualities (contrast, uniformity).
  • Significantly same structural measurements between DeepRecon and NADR images.
• ACS:
  • Equivalence in performance to ACS on the uMR Omega (K220332).
  • Better performance than CS in SNR and resolution.
  • Maintenance of image qualities (contrast, uniformity) compared to fully sampled data (golden standard).
  • Significantly same structural measurements between ACS and fully sampled images.

Table of Inferred Acceptance Criteria and Reported Device Performance:

Breakdown of the Study as described in the 510(k) Summary:

2. Sample size used for the test set and the data provenance:

• DeepRecon:

  • "The testing dataset for performance testing was collected independently from the training dataset, with separated subjects and during different time periods."
  • The exact sample size (number of subjects/cases) for the DeepRecon test set is not specified beyond being "independent."
  • Data Provenance: Implied to be from UIH MRI systems, likely from clinical or volunteer scans. No specific country of origin or retrospective/prospective nature is stated for the test datasets, but training data was "collected from 264 volunteers" and "165,837 cases" using "UIH MRI systems," which suggests internal company data, likely from China where the company is based. The testing data is independently collected.

• ACS:

  • "The training and test datasets are collected from 35 volunteers, including 24 males and 11 females, ages ranging from 18 to 60. The samples from these volunteers are distributed randomly into training and test datasets."
  • "The validation dataset is collected from 15 volunteers, including 10 males and 5 females, whose ages range from 18 to 60."
  • It specifies "35 volunteers" for training+test and "15 volunteers" for validation. The text states "testing dataset for performance testing was collected independently from the training dataset," which contradicts the "distributed randomly into training and test datasets" statement for the 35 volunteers. This requires clarification, but assuming the 35 volunteers contributed to both, the total number used for testing is not explicitly broken out from the 35. The "validation dataset" of 15 volunteers seems to be an additional independent test set.
  • Data Provenance: Implied to be from UIH MRI systems. No specific country of origin or retrospective/prospective nature is stated.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• Expert Review: "Sample clinical images for all clinical sequences and coils were reviewed by U.S. board-certified radiologist comparing the proposed device and predicate device."
  • Number of experts: Not specified, only "radiologist" (singular or plural not clear).
  • Qualifications: "U.S. board-certified radiologist." No mention of years of experience.
• Quantitative/Objective Ground Truth: For DeepRecon and ACS, ground truth was not established by experts but rather by specific technical methods:
  • DeepRecon: "multiple-averaged images with high-resolution and high SNR were collected as the ground-truth images."
  • ACS: "Fully-sampled k-space data were collected and transformed to image space as the ground-truth."

4. Adjudication method (e.g., 2+1, 3+1, none) for the test set:

• The document implies a technical assessment for AI performance (SNR, resolution, structural measurements). For the "U.S. board-certified radiologist" review, no specific adjudication method (e.g., 2+1 consensus) is mentioned.

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:

• No MRMC comparative effectiveness study involving human readers and AI assistance is described. The performance evaluation focuses on the technical imaging characteristics and comparison to the predicate device or baseline (NADR/CS). The "U.S. board-certified radiologist" review seems to be a qualitative assessment of diagnostic image quality rather than a structured MRMC study with quantitative outcomes.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:

• Yes, the performance tests for DeepRecon and ACS are described as standalone evaluations of the algorithms' effects on image quality (SNR, resolution, contrast, uniformity, structural measurements) by comparing them to NA (No Algorithm) or baseline (CS) methods.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

• DeepRecon: "multiple-averaged images with high-resolution and high SNR" (objective, technical ground truth representing optimal image quality).
• ACS: "Fully-sampled k-space data" (objective, technical ground truth representing complete data).
• For the qualitative review by the radiologist, the "diagnostic quality images" from the predicate device implicitly served as a reference or ground truth for comparison.

8. The sample size for the training set:

• DeepRecon: "264 volunteers" resulting in "165,837 cases."
• ACS: "35 volunteers" (randomly distributed into training and test datasets). The exact split for training is not specified but is part of this 35.

9. How the ground truth for the training set was established:

• DeepRecon: "the multiple-averaged images with high-resolution and high SNR were collected as the ground-truth images." "All data were manually quality controlled before included for training."
• ACS: "Fully-sampled k-space data were collected and transformed to image space as the ground-truth." "All data were manually quality controlled before included for training."

In summary, the provided document focuses on demonstrating technical equivalence and improved image characteristics for the AI components (DeepRecon, ACS) through non-clinical testing against technically derived ground truths, rather than a clinical multi-reader study with expert consensus ground truth or outcomes data. The human reader involvement seems to be a qualitative review of diagnostic image quality rather than a formal MRMC study.

Ask a specific question about this device

uOmnispace.MR is a software solution intended to be used for viewing, manipulating and analyzing medical images. It supports interpretation and evaluation of examinations within healthcare institutions. It has the following additional indications:

The uOmnispace.MR Stitching is intended to create full-format images from overlapping MR volume data sets acquired at multiple stages.

The uOmnispace.MR Dynamic application is intended to provide a general postprocessing tool for time course studies.

The uOmnispace.MR MRS (MR Spectroscopy) is intended to evaluate the molecule constitution and spatial distribution of cell metabolism. It provides a set of tools to view, process, and analyze the complex MRS data. This application supports the analysis for both SVS (Single Voxel Spectroscopy) and CSI (Chemical Shift Imaging) data.

The uOmnispace.MR MAPs application is intended to provide a number of arithmetic and statistical functions for evaluating dynamic processes and images. These functions are applied to the grayscale values of medical images.

The uOmnispace.MR Breast Evaluation application provides the user a tool to calculate parameter maps from contrast-enhanced time-course images.

The uOmnispace.MR Brain Perfusion application is intended to allow the visualization of temporal variations in the dynamic susceptibility time series of MR datasets.

· MR uOmnispace.MR Vessel Analysis is intended to provide a tool for viewing, manipulating, and evaluating MR vascular images.

The uOmnispace.MR DCE analysis is intended to view, manipulate, and evaluate dynamic contrast-enhanced MRI images.

The uOmnispace.MR United Neuro is intended to view, manipulate MR neurological images.

■ The uOmnispace.MR Cardiac Function is intended to view, evaluate functional analysis of cardiac MR images.

The uOmnispace.MR Flow Analysis is intended to view, evaluate flow analysis of flow MR images.

The uOmnispace.MR is a post-processing software based on the uOmnispace platform (cleared in K230039) for viewing, manipulating, evaluating and analyzing MR images, can run alone or with other advanced commercially cleared applications.

This proposed device contains the following applications:

• uOmnispace.MR Stitching
• uOmnispace.MR Dynamic
• uOmnispace.MR MRS
• uOmnispace.MR MAPs
• uOmnispace.MR Breast Evaluation
• . uOmnispace.MR Brain Perfusion
• uOmnispace.MR Vessel Analysis
• uOmnispace.MR DCE Analysis
• uOmnispace.MR United Neuro
• uOmnispace.MR Cardiac Analysis
• uOmnispace.MR Flow Analysis

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: 114 samples from 114 different patients.
• Data Provenance: The data includes patients of various genders (35 Male, 20 Female, 59 Unknown), ages (5 between 14-25, 12 between 25-40, 22 between 40-60, 13 between 60-79, 62 Unknown), and ethnicities (50 Europe, 53 Asia, 11 USA). The data was acquired using MR scanners from various manufacturers: UIH (58), GE (2), Philips (2), Siemens (52), and with different magnetic field strengths: 1.5T (23), 3.0T (41), 50 Unknown. The text does not explicitly state if the data was retrospective or prospective, but the mention of a "deep learning-based Automatic ventricular segmentation Algorithm for the LV&RV Contour Segmentation feature" and "The performance testing for deep learning-based Automatic ventricular segmentation Algorithm was performed on 114 subjects...during the product development" implies a retrospective study using existing data to validate the developed algorithm.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

The test set's ground truth was established by comparing the proposed device's results with those of the predicate device. The text does not explicitly state that human experts established the ground truth for the test set by manually segmenting the images for direct comparison against the algorithm's output. Instead, it seems the predicate device's output serves as the "ground truth" for the comparison of the new device's algorithm.

However, for the training ground truth, the following was stated:

• Number of Experts: Two cardiologists.
• Qualifications: Both cardiologists had "more than 10 years of experience each."

4. Adjudication Method for the Test Set

The study does not describe an adjudication method for the test set in the conventional sense of multiple human readers independently assessing the cases. Instead, the comparison is made between the proposed device's algorithm output and the predicate device's output.

For the training ground truth, the following adjudication method was used:

• Manual tracing was performed by an experienced user.
• Validation of these contours was done by two independent experts (more than 10 years experience).
• If there was a disagreement, a consensus between the experts was reached.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size

No MRMC comparative effectiveness study was done to assess how much human readers improve with AI vs without AI assistance. The study focuses on comparing the proposed device's algorithm performance directly against a predicate device's cardiac function application based on the Dice coefficient.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Yes, a standalone performance study was done for the "deep learning-based Automatic ventricular segmentation Algorithm" for the LV&RV Contour Segmentation feature. The device's algorithm output was directly compared to the output of the predicate device's cardiac function application using the Dice coefficient.

7. The Type of Ground Truth Used

For the test set, the "ground truth" for comparison was the output of the cardiac function application of the predicate device.

For the training set, the ground truth was expert consensus based on manual tracing by an experienced user and validated by two independent cardiologists with over 10 years of experience.

8. The Sample Size for the Training Set

The document states: "The training data used for the training of the cardiac ventricular segmentation algorithm is independent of the data used to test the algorithm." However, it does not provide the specific sample size for the training set.

9. How the Ground Truth for the Training Set Was Established

The ground truth for the training set was established through manual annotation and expert consensus:

• It was "manually drawn on short axis slices in diastole and systole by two cardiologists with more than 10 years of experience each."
• "Manual tracing of the cardiac was performed by an experienced user."
• "The validation of these contours was done by two independent expert (more than 10 years) in this domain."
• "If there is a disagreement, a consensus between the experts was done."

Ask a specific question about this device

The uMR 680 system is indicated for use as a magnetic resonance diagnostic device (MRDD) that produces sagittal, transverse, coronal, and oblique cross sectional images, and spectroscopic images, and that display internal anatomical structure and/or function of the head, body and extremities. These images and the physical parameters derived from the images when interpreted by a trained physician yield information that may assist the diagnosis. Contrast agents may be used depending on the region of interest of the scan.

The uMR 680 is a 1.5T superconducting magnetic resonance diagnostic device with a 70cm size patient bore. It consists of components such as magnet, RF power amplifier, RF coils, gradient power amplifier, gradient coils, patient table, spectrometer, computer, equipment cabinets, power distribution system, internal communication system, and vital signal module etc. The uMR 680 Magnetic Resonance Diagnostic Device is designed to conform to NEMA and DICOM standards. uMR 680 has been previously cleared by FDA via K222755. The modification performed on uMR 680 (K222755) in this submission is due to the addition of - Breast Coil -24 - epi_se_mre - MRE (Magnetic Resonance Elastography) The modification does not affect the intended use or alter the fundamental scientific technology of the device.

The provided text describes the acceptance criteria and the study results for the Shanghai United Imaging Healthcare Co., Ltd. uMR 680 Magnetic Resonance Diagnostic Device.

Here's the breakdown of the information requested:

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample size used for the test set and the data provenance:

• Sample Size: Not explicitly stated for each test beyond "phantom and volunteer test" for MRE performance and "the image generated by Breast Coil-24" for clinical image quality.
• Data Provenance: Not specified. It does not mention country of origin or if the data was retrospective or prospective.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• Number of Experts: For clinical image quality for both Breast Coil - 24 and MRE, it states "The U.S. Board Certified radiologist approves." This implies at least one, but the exact number of radiologists is not specified.
• Qualifications of Experts: "U.S. Board Certified radiologist."

4. Adjudication method for the test set:

• Adjudication Method: Not explicitly stated. The statement "The U.S. Board Certified radiologist approves" suggests a single expert's approval for clinical image quality assessment, rather than a multi-expert consensus method like 2+1 or 3+1.

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 explicitly mentioned as being done for this submission. The studies detailed focus on system performance parameters and clinical image quality approval by a radiologist.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:

• This device is a Magnetic Resonance Diagnostic Device (MRDD), which is a hardware system that produces images and physical parameters to be interpreted by a trained physician. The performance tests described (surface heating, electrical/mechanical safety, SNR, uniformity, biocompatibility, EMC, performance, and clinical image quality) are for the device itself and its components/functionalities, not for an AI algorithm working in a standalone capacity. So, this question is not directly applicable in the terms of an AI algorithm, but the system performance was evaluated standalone from human interpretation in most of the tests.

7. The type of ground truth used:

• Clinical Image Quality: Expert consensus (or approval by a U.S. Board Certified radiologist) that the image quality is sufficient for diagnostic use.
• Biocompatibility: Demonstrated by using materials identical to a previously cleared device (uMR Omega, K230152) and meeting regulatory guidance.
• Other performance metrics (Surface heating, electrical/mechanical safety, SNR, Uniformity, EMC, MRE Performance): Based on compliance with established engineering and safety standards (NEMA MS 14, ANSI/AAMI ES60601-1, NEMA MS 1, NEMA MS 3, NEMA MS 6, NEMA MS 9, IEC 60601-1-2, IEC 60601-4-2) and design specifications.

8. The sample size for the training set:

• The document does not detail any "training set" as it is describing a hardware medical device with specific new coils and pulse sequences rather than a machine learning or AI-driven diagnostic algorithm that would typically require a training set.

9. How the ground truth for the training set was established:

• Not applicable, as no training set for an AI/ML algorithm is mentioned.

Ask a specific question about this device