Search Results

The Swoop Portable MR Imaging System is a portable, ultra-low field magnetic resonance imaging device for producing images that display the internal structure of the head where full diagnostic examination is not clinically practical. When interpreted by a trained physician, these images provide information that can be useful in determining a diagnosis.

The Swoop system is portable, ultra-low field MRI device that enables visualization of the internal structures of the head using standard magnetic resonance imaging contrasts. The main interface is a commercial off-the-shelf device that is used for operating the system, providing access to patient data, exam setup, exam execution, viewing MRI image data for quality control purposes, and cloud storage interactions. The system can generate MRI data sets with a broad range of contrasts. The Swoop system user interface includes touch screen menus, controls, indicators, and navigation icons that allow the operator to control the system and to view imagery. The Swoop System image reconstruction algorithm utilizes deep learning to provide improved image quality for T1W, T2W, and FLAIR sequences, specifically in terms of reductions in image noise and blurring.

Here's an analysis of the provided text to extract information about the acceptance criteria and the study that proves the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't present a direct table of acceptance criteria versus reported performance in the typical sense of numerical metrics for a specific clinical task (e.g., sensitivity, specificity for disease detection). Instead, the performance is demonstrated through various non-clinical tests and compliance with established standards.

Summary of Device Features (where performance is implicit):

• Image Reconstruction Algorithm: Utilizes deep learning for improved image quality (reductions in image noise and blurring) for T1W, T2W, and FLAIR sequences.
• DWI Image Post-processing: Fast Iterative Shrinkage Thresholding Algorithm (FISTA).
• Image Post-Processing (General): Advanced Denoising (T1W, T2W, FLAIR, DWI), image orientation transform, geometric distortion correction, receive coil intensity correction, DICOM output.

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

The document does not detail a "test set" in the context of a clinical study with patient data. The evaluation is primarily based on non-clinical performance testing and software verification/validation. Therefore, there is no specific sample size of patients or images from a clinical test set mentioned.

The data provenance for the non-clinical tests would be from internal Hyperfine labs or certified testing facilities that perform imaging phantom tests and software testing.

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

Not applicable in the context of this submission. The "ground truth" for the non-clinical tests is established by industry standards (NEMA, ACR phantom guidelines, ISO, etc.) and engineering requirements rather than expert clinical consensus on patient data.

4. Adjudication Method for the Test Set

Not applicable. There was no clinical test set requiring expert adjudication.

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 is mentioned in the provided text. The document refers to "deep learning to provide improved image quality" for certain sequences, but this is an inherent part of the device's image reconstruction algorithm and not a specific AI-assisted diagnostic aid for human readers. The clinical utility of these improved images (once interpreted by a trained physician) is part of the overall "diagnosis," but no study on AI assistance for human readers is described here.

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

The device is an imaging system; its output is the image, which then requires interpretation by a trained physician. The deep learning component is integrated into the image reconstruction, affecting the quality of the image produced by the algorithm. Therefore, in a sense, the "standalone" performance of the image generation (algorithm only) is what's being evaluated against image quality criteria through non-clinical means. There isn't a separate "AI algorithm" that outputs a diagnostic decision independently.

7. The Type of Ground Truth Used

The ground truth used for demonstrating compliance is primarily:

• Engineering Requirements and Standards: For software functionalities, electrical safety, biocompatibility, cleaning/disinfection, and cybersecurity.
• Imaging Phantoms: For image quality criteria, based on established standards like NEMA and ACR phantom test guidance.

8. The Sample Size for the Training Set

The document states that the image reconstruction algorithm "utilizes deep learning." However, it does not provide any information on the sample size (number of images, patients, etc.) for the training set used for this deep learning model.

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

The document does not provide details on how the ground truth for the deep learning training set was established. It only mentions that deep learning is used for image reconstruction to improve image quality (noise reduction and blurring). For image reconstruction deep learning models, the "ground truth" during training typically involves pairs of lower-quality input images and corresponding higher-quality reference images (often obtained using conventional non-accelerated MRI or higher field strength MRI) or synthetic data representing ideal image characteristics. However, specifics are not in this text.

Ask a specific question about this device