Search Results

The ECHELON Synergy System is an imaging device and is intended to provide the physician with physiological and clinical information, obtained non-invasively and without the use of ionizing radiation. The MR system produces transverse, coronal, sagittal, oblique, and curved cross sectional images that display the internal structure of the head, body, or extremities. The images produced by the MR system reflect the spatial distribution of protons (hydrogen nuclei) exhibiting magnetic resonance. The NMR properties that determine the image appearance are proton density, spinlattice relaxation time (TI), spin-spin relaxation time (T2) and flow. When interpreted by a trained physician, these images provide information that can be useful in diagnosis determination.

Anatomical Region: Head, Body, Spine, Extremities Nucleus excited: Proton

Diagnostic uses:

• · TI, T2, proton density weighted imaging
• · Diffusion weighted imaging
• · MR Angiography
• · Image processing
• · Spectroscopy
• · Whole Body

The ECHELON Synergy is a Magnetic Resonance Imaging System that utilizes a 1.5 Tesla superconducting magnet in a gantry design. Magnetic Resonance imaging (MRI) is based on the fact that certain atomic nuclei have electromagnetic properties that cause them to act as small spinning bar magnets. The most ubiquitous of these nuclei is hydrogen, which makes it the primary nuclei currently used in magnetic resonance imaging. When placed in a static maqnetic field, these nuclei assume a net orientation or alignment with the magnetic field, referred to as a net magnetization vector. The introduction of a short burst of radiofrequency (RF) excitation of a wavelength specific to the magnetic field strength and to the atomic nuclei under consideration can cause a re-orientation of the net magnetization vector. When the RF excitation is removed, the protons relax and return to their original vector. The rate of relaxation is exponential and varies with the character of the proton and its adjacent molecular environment. This re-orientation process is characterized by two exponential relaxation times, called T1 and T2. A RF emission or echo that can be measured accompanies these relaxation events. The emissions are used to develop a representation of the relaxation events in a three dimensional matrix. Spatial localization is encoded into the echoes by varving the RF excitation. applying appropriate magnetic field gradients in the x, y, and z directions, and changing the direction and strength of these gradients. Images depicting the spatial distribution of the NMR characteristics can be reconstructed by using image processing techniques similar to those used in computed tomography.

The provided document describes the Fujifilm ECHELON Synergy V10.0 MRI system, which is an updated version of a previously cleared device. The submission focuses on demonstrating substantial equivalence to the predicate device (ECHELON Synergy MRI System K223426) by highlighting changes and providing performance evaluations.

Here's an analysis of the acceptance criteria and study information:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a table of "acceptance criteria" for the overall device in a quantifiable format. Instead, it demonstrates the new features' performance through clinical image testing and phantom studies, comparing them to conventional methods or manual positioning. The acceptance criteria for "DLR Clear" are implied through achieving statistical significance for superiority in certain image quality metrics over conventional imaging and clinical acceptability. For "AutoPose," the criteria are implied through reduction or equivalence in time and steps for slice positioning.

Here's a summary of the performance results for the new features (DLR Clear and AutoPose):

2. Sample Sizes and Data Provenance

DLR Clear:

• Test Set Sample Size: 53 unique subjects (patients and volunteers).
• Data Provenance: U.S. and Japan.
• Retrospective/Prospective: Not explicitly stated, but the description "scanned in the anatomical regions... for the test datasets separately from the training and validation datasets" suggests a prospective collection for the test set or careful selection from retrospective archives to act as a distinct test set.

AutoPose (Spine, Breast, HipJoint):

• Test Set Sample Size:
  • Spine: 177 cases
  • Breast: 66 cases
  • HipJoint: 65 cases
• Data Provenance: FUJIFILM Healthcare Corporation and clinical site (country not specified for clinical site, but assuming Japan given the company origin and DLR Clear's data provenance).
• Retrospective/Prospective: Not explicitly stated.

3. Number of Experts and Qualifications for Ground Truth for the Test Set

DLR Clear:

• Number of Experts: Three (3) US board-certified radiologists.
• Qualifications of Experts: US board-certified radiologists. (Specific experience levels not provided).

AutoPose:

• Number of Experts: "Certified radiological technologists." (Number not specified, specific experience not provided). Their evaluation was on the reduction/equivalence of time and steps for slice positioning.

4. Adjudication Method for the Test Set

DLR Clear:

• The images were randomized and blinded to the reviewers.
• Reviewers compared image quality metrics (truncation artifact reduction, image sharpness, lesion conspicuity, and overall image quality) using a 3-point scale.
• The results indicate statistical analysis (p<0.05), implying that the individual ratings were aggregated or compared statistically to determine superiority. It does not explicitly state a consensus-based adjudication method (like 2+1 or 3+1). It seems individual reviewer scores were used.

AutoPose:

• Evaluated by "certified radiological technologists." The method of combining their evaluations or achieving consensus is not explicitly described. It states "They evaluated that almost cases..." suggesting a collective assessment or majority opinion, but no formal adjudication process is detailed.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

There is no explicit MRMC comparative effectiveness study directly comparing human reader performance with and without AI assistance (i.e., effect size of human improvement) described in this document.

The study for DLR Clear involved radiologists reviewing images reconstructed with DLR Clear versus conventional methods. This evaluates the impact of the AI-powered reconstruction on image quality, which in turn facilitates human reading, but it does not measure the human reader's diagnostic performance improvement (e.g., AUC, sensitivity, specificity) when using the AI-processed images compared to conventional images.

6. Standalone Performance Study

DLR Clear:
Yes, a form of standalone performance was done for DLR Clear. The phantom testing ("Total Validation, Relative Edge Sharpness, and FWHM") demonstrates the algorithm's direct performance on objective metrics before human interpretation. The clinical image review also evaluates the quality of the algorithm's output directly, without explicitly involving an "AI-assisted reading" scenario. The radiologists were evaluating the images produced by the algorithm, not acting as diagnosticians making decisions with or without AI output.

AutoPose:
Yes, the AutoPose feature's performance (reduction in time/steps for positioning) was evaluated by radiological technologists, which can be considered a standalone performance assessment for the positioning automation.

7. Type of Ground Truth Used

DLR Clear:

• Phantom Testing: The ground truth for phantom testing is the physical properties of the cylindrical and ACR phantoms themselves, measured by objective metrics (e.g., FWHM for spatial resolution).
• Clinical Testing: The ground truth for clinical image quality assessment relies on expert consensus/opinion (ratings by 3 US board-certified radiologists) regarding image quality metrics (truncation artifact, sharpness, lesion conspicuity, overall image quality) and clinical acceptability. There's no mention of pathology or outcomes data as ground truth for these image quality assessments.

AutoPose:

• The ground truth for AutoPose is implicitly the objective measurement of time and steps required for slice positioning, compared against manual methods, as evaluated by certified radiological technologists.

8. Sample Size for the Training Set

DLR Clear: Not explicitly stated. The document mentions "test datasets separately from the training and validation datasets" for DLR Clear, indicating that training data existed, but its size is not provided.

AutoPose: Not explicitly stated.

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

DLR Clear: Not explicitly stated. Since it's a deep learning reconstruction (DLR), the ground truth for training would typically involve pairs of raw MRI data and high-quality, potentially "gold standard" reconstructed images (e.g., using more extensive acquisition or advanced reconstruction techniques) or specific labels for image quality characteristics.

AutoPose: Not explicitly stated. For an automated positioning algorithm, training ground truth would likely involve expert-defined optimal slice prescriptions on a large dataset of diverse anatomies.

Ask a specific question about this device

The OASIS MRI System is an imaging device, and is intended to provide the physician with physiological and clinical information, obtained non-invasively and without the use of ionizing radiation. The MR system produces transverse, coronal, sagittal, oblique, and curved crosssectional images that display the internal structure of the head, body, or extremities. The images produced by the MR system reflect the spatial distribution of protons (hydrogen nuclei) exhibiting magnetic resonance. The NMR properties that determine the image appearance are proton density, spin-lattice relaxation time (T1), spin-spin relaxation time (T2), and flow. When interpreted by a trained physician. these images provide information that can be useful in diagnosis determination.

Anatomical Region: Head, Body, Spine, Extremities
Nucleus excited: Proton
Diagnostic uses: T1, T2, proton density weighted imaging
Diffusion weighted imaging
MR Angiography
Image processing
Spectroscopy
Whole Body

The OASIS MRI System is a Magnetic Resonance Imaging System that utilizes a 1.2 Tesla superconducting magnet in a gantry design. Magnetic Resonance imaging (MRI) is based on the fact that certain atomic nuclei have electromagnetic properties that cause them to act as small spinning bar magnets. The most ubiquitous of these nuclei is hydrogen, which makes it the primary nuclei currently used in magnetic resonance imaging. When placed in a static magnetic field, these nuclei assume a net orientation or alignment with the magnetic field, referred to as a net magnetization vector. The introduction of a short burst of radiofrequency (RF) excitation of a wavelength specific to the magnetic field strength and to the atomic nuclei under consideration can cause a re-orientation of the net magnetization vector. When the RF excitation is removed, the protons relax and return to their original vector. The rate of relaxation is exponential and varies with the character of the proton and its adjacent molecular environment. This re-orientation process is characterized by two exponential relaxation times, called T1 and T2. A RF emission or echo that can be measured accompanies these relaxation events.

The emissions are used to develop a representation of the relaxation events in a three dimensional matrix. Spatial localization is encoded into the RF excitation, applying appropriate magnetic field gradients in the x, y, and z directions, and changing the direction and strength of these gradients. Images depicting the spatial distribution of the NMR characteristics can be reconstructed by using image processing techniques similar to those used in computed tomography.

The provided document, a 510(k) Summary for the OASIS MRI System (K240571), describes the device and its equivalence to a predicate device (OASIS MRI System K211406). The acceptance criteria and performance study details are primarily focused on a new feature, DLR Rise.

Here's a breakdown of the requested information:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state numerical acceptance criteria for DLR Rise. Instead, the acceptance is based on expert subjective evaluation of image quality metrics and clinical acceptability.

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

• Sample Size: 71 unique subjects (patients and volunteers).
• Data Provenance: Imaging data collected from the U.S. and Japan. The study was retrospective, using pre-existing data or data collected specifically for this evaluation. (The text states "scanned in the anatomical regions... to provide the test datasets separately from the training and validation datasets," implying these were fresh scans for testing purposes, but the retrospective/prospective nature isn't explicitly stated for the overall study design.)

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

• Number of Experts: Three (3) US board-certified radiologists.
• Qualifications: Each radiologist had at least 5 years' experience since residency.

4. Adjudication Method for the Test Set

• Adjudication Method: Majority decision. After each radiologist reviewed the images, the combination of the three radiologists' answers was obtained by majority decision, and the combined answers were used for the evaluation.

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

• Yes, a multi-reader, multi-case study was performed.
• Effect Size of Human Readers Improvement with AI vs. Without AI Assistance: The study directly compared images reconstructed with DLR Rise (AI-assisted) against conventional images (without DLR Rise/AI-assisted). The results indicated that for SNR, sharpness, lesion conspicuity, and overall image quality, the images with DLR Rise were superior with a statistically significant difference (p<0.05). Furthermore, all images with DLR Rise were rated as "better or equivalent" compared to images without DLR Rise, and all were clinically acceptable. This indicates a significant positive effect on image quality, which directly aids human readers. The exact quantitative effect size (e.g., AUC difference, percentage improvement) is not provided beyond the statistical significance and qualitative superiority.

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

• No, the study design explicitly involved human readers (radiologists) reviewing and comparing the images. It was not a standalone algorithm performance evaluation. The DLR Rise is an image processing feature, not a diagnostic AI algorithm that provides direct diagnostic outputs without human interpretation.

7. The Type of Ground Truth Used

• Ground Truth Type: Expert consensus based on subjective evaluation of image quality metrics and clinical acceptability by three board-certified radiologists using a majority decision. There is no mention of pathology or outcomes data being used as ground truth for this image quality assessment.

8. The Sample Size for the Training Set

• The document states that the test datasets were provided "separately from the training and validation datasets." However, the sample size for the training set is not explicitly provided in the excerpt. It mentions that "The AI algorithm and learning condition of DLR Rise in the subject device are the same as those of DLR Rise in the reference device (K223426)." This implies the training was done for the DLR Rise algorithm itself, but the sample size for that training is not detailed in this document.

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

• The document does not provide details on how the ground truth for the training set was established. It only states that the AI algorithm and its learning conditions for DLR Rise are the same as those in the reference device (K223426). This suggests that the ground truth establishment for training would have been part of the development/validation for the DLR Rise feature that was included in the reference device, but the specifics are not included here.

Ask a specific question about this device