Search Results

• · is a software option intended for use on Achieva and Ingenia 1.5T & 3.0T MR Systems
• · is a non-invasive triglyceride fat fraction calculation
• · is a 3D, single breath-hold acquisition and reconstruction technique
• · allows for water-fat separation generating Water-only, Fat-only images as well as Out-phase and In-phase images
• · is indicated for magnetic resonance imaging of the liver

mDIXON-Quant acquisition and reconstruction is based on the mDIXON product developed previously. Acquisition relies on a gradient echo acquisition (FFE) including a large number of echoes (6 or more). Furthermore, mDIXON-Quant reconstruction includes the use of multiple spectral peaks of triglyceride fat, correction of the T2* confounding effect and reduction of T1 bias.

mDIXON-Quant allows for water-fat separation and generates Water-only, Fat-only images as well as Out-phase and In-phase images synthesized from the Water and Fat images. Additionally, mDIXON-Ouant produces images representing triglyceride fat fraction as well as images representing transverse magnetization relaxation.

The feature requires:

• · Specific parameter settings for the mDIXON sequence, within cleared parameter limits, to acquire the MR signals
• · A new calculation function to generate the new images for Fat Fraction, T2 *. This function uses a set of MR images as input that is generated in a cleared manner from the acquired MR signals.
• · The new images need to be stored and displayed with the appropriate labels for Fat Fraction, T2 *, applying the facilities provided by the cleared platform.

Here's a summary of the acceptance criteria and study information based on the provided text for the mDIXON-Quant device:

1. Table of Acceptance Criteria and Reported Device Performance

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

The document does not explicitly state a specific "test set" sample size in terms of number of cases or patients for a clinical validation study. The clinical validation involved "volunteers."

• Sample Size (clinical validation): Not explicitly stated, but involved "all volunteers."
• Data Provenance: The document does not specify the country of origin of the data. The clinical validation was prospective ("The clinical user needs were tested as part of the validation testing... All volunteers were able to hold their breath for that time. As such, the clinical validation of mDIXON-Quant completed successfully.").

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

The document does not mention the use of experts to establish ground truth for a test set, nor does it detail any qualifications for such experts. The primary assessments for the clinical validation appear to be directly related to the device's functional output (image generation, ROI functionality, etc.) and user experience (breath-hold, workflow smoothness).

4. Adjudication Method for the Test Set

Not applicable. The document does not describe any expert adjudication process for establishing ground truth or evaluating disagreements, as it primarily focuses on device functionality and user experience during clinical validation.

5. If a Multi Reader Multi Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size of How Much Human Readers Improve with AI vs Without AI Assistance

No, a Multi Reader Multi Case (MRMC) comparative effectiveness study was not reported. The study focused on the performance of the mDIXON-Quant device itself, not on how human readers' diagnostic accuracy is affected by using the device. The device is a "non-invasive triglyceride fat fraction calculation" and "generates Water-only, Fat-only images as well as Out-phase and In-phase images," implying it provides quantitative data and separated images, not an AI-assisted diagnostic classification.

6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done

Yes, a standalone performance assessment was conducted for the algorithm's core functionality, particularly for quantification accuracy and reproducibility:

• Quantification Performance testing: "test results showed that the FF accuracy is ±3.5% for both field strengths and all possible parameter combinations, and the T2* range and reproducibility is 1.5% for both field strengths."
• Acquisition and Reconstruction testing: Verified "FF and T2* images were available."

These tests demonstrate the algorithm's ability to produce accurate and reproducible quantitative results and generate the intended image types independently of human interpretation of those results.

7. The Type of Ground Truth Used

The ground truth for the quantitative performance (Fat Fraction accuracy, T2* range and reproducibility) was established against an assumed reference standard for these measurements, likely using phantoms or controlled experimental conditions where the true values were known. The document states "Quantification Performance testing" but does not explicitly name the reference method for establishing this ground truth. For the qualitative aspects of clinical validation (e.g., image availability, ROI capability, workflow), the ground truth was essentially whether the device successfully performed the intended functions as observed by the testers.

8. The Sample Size for the Training Set

The document does not provide any information regarding a training set or its sample size. This is a medical imaging reconstruction and quantification software, not a machine learning model that typically requires a distinct training set in the conventional sense. The "training" of such a system would involve engineering and calibration against known physical properties and signal characteristics rather than labeled clinical data.

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

Not applicable, as no training set (in the context of machine learning) is described. The development of such an algorithm would rely on established physics and signal processing principles, validated against phantom studies and potentially pre-existing clinical data for fine-tuning.

Ask a specific question about this device