Search Results

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 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, 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.

The OASIS is a Magnetic Resonance Imaging System that utilizes a 1.2 Tesla superconducting maqnet in a qantry design.

After reviewing the provided text, it's immediately clear that this document is a 510(k) Premarket Notification from the FDA for a new version of an MRI system (OASIS MRI System).

The key takeaway is that this document DOES NOT describe a study involving an "AI" or "algorithm" with human-in-the-loop performance, nor does it provide detailed quantitative acceptance criteria and study results for such a device.

Instead, it's a submission for the OASIS MRI System w/ V6.0F being substantially equivalent to its predicate device, the OASIS MRI System (K093044). The "performance evaluation" mentioned focuses on demonstrating that new features (like updated coils, operating systems, and software versions) do not negatively impact the device's safety and effectiveness and that the system performs as intended for diagnostic use.

The "performance evaluation" is primarily rational analysis and bench testing, with "clinical image examples" provided for usability, rather than a rigorous clinical performance study with defined quantitative acceptance criteria. There's no mention of:

• Specific quantitative metrics like sensitivity, specificity, AUC.
• Sample sizes for test sets in the typical sense of a clinical trial.
• Ground truth establishment by experts for specific diagnostic performance.
• MRMC studies or human reader improvement with AI assistance.
• Standalone algorithm performance.

Therefore, I cannot fulfill most of your request directly from the provided text, as the information regarding the acceptance criteria and the study design for an AI/algorithm-based device is not present. The document focuses on demonstrating substantial equivalence of a general MRI system, not the performance of an AI application within it.

I can, however, extract information about the predicate device and the general approach to demonstrating equivalence, which might give you some insight into the FDA 510(k) process for MRI systems, but not for an AI/algorithm.

Based on the provided document, here's what can be extracted, and where limitations exist regarding your specific request:

Limitations: The document does not describe the acceptance criteria and a study for an AI/algorithm-based device. It is a 510(k) submission for a new version of an MRI system (hardware and software updates) demonstrating substantial equivalence to a legally marketed predicate device. Therefore, many of the requested points related to AI/algorithm performance, ground truth establishment, expert readers, and MRMC studies are not applicable or not detailed in this context.

Information Extracted (with caveats for non-AI focus):

1. A table of acceptance criteria and the reported device performance:

   • Acceptance Criteria (Implied for Substantial Equivalence): The primary "acceptance criterion" for this 510(k) is that the modified OASIS MRI System (V6.0F) is substantially equivalent to its predicate (OASIS K093044) in terms of safety, effectiveness, and intended use, despite specific updates in coils, operating system, CPU, and application software. This is demonstrated by showing that new features "perform as intended for diagnostic use" and that differences do not "raise different questions of safety and effectiveness."
   • Reported Device Performance: The "performance" is qualitative, focusing on whether new features function correctly and that fundamental safety/performance characteristics (like signal-to-noise ratio, uniformity, acoustic noise, electrical safety, EMC) remain acceptable or are not negatively impacted.
   • Table 1: Performance Analysis

     Testing Type	Rationale Analysis	Reported Device Performance
     Performance Testing - Bench	Performance bench testing was conducted on the applicable new features.	Test data confirmed that each new feature perform as intended for diagnostic use.
     Performance Testing - Clinical	Clinical image examples are provided for each applicable new feature and that we judged to be sufficient to evaluate clinical usability.	[Details of usability are not quantified, but the judgment was "sufficient"]

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

   • The document mentions "clinical image examples" for usability but does not specify a sample size for a clinical test set in the way one would for an AI performance study.
   • Data Provenance: Not specified. The clinical images are "examples" and likely collected by Hitachi, but whether they are retrospective or prospective, or from specific countries, is not stated.

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

   • Not applicable directly. This document is for an MRI system, not an AI/algorithm that requires expert-established ground truth for diagnostic accuracy. The "clinical image examples" were "judged to be sufficient to evaluate clinical usability," which implies interpretation by presumably qualified personnel (likely radiologists or technologists), but the number and qualifications are not specified nor is there a formal "ground truth" establishment process described for a test set.

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

   • Not applicable. There is no formal adjudication method described for a test set, as this is not a study assessing diagnostic performance of an algorithm.

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, an MRMC comparative effectiveness study was NOT done. This document does not describe a study involving AI assistance for human readers.

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

   • No, a standalone algorithm performance study was NOT done. This document describes a medical imaging device (MRI system), not an AI algorithm.

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

   • Not applicable in the context of an AI algorithm. For the MRI system itself, the "ground truth" for demonstrating substantial equivalence relies on established industry standards (NEMA, IEC) for image quality, safety parameters (e.g., acoustic noise, SAR), and the system's ability to produce images useful for diagnosis, interpreted by a "trained physician". This is not a ground truth for a specific diagnostic outcome.

8. The sample size for the training set:

   • Not applicable. This document describes an MRI system, not an AI model requiring a training set. The changes are primarily software version updates and new coils for an existing hardware platform.

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

   • Not applicable. See point 8.

Ask a specific question about this device