Search Results

The HI STAR system is indicated for use a general purpose MR imager that produces transverse, sagittal, coronal, and oblique cross-sectional images of internal structures of the human body, head, and extremities. Contrast in these images reflect the intrinsic oroperties of the hydrogen proton, proton density, T1, and T2, and the extrinsic parameters of the applied imaging sequences. The indications for the HI STAR do not differ significantly from the predicate, or legally marketed devices.

The Health Images HI STAR is a proton (H-1) Magnetic Resonance Imaging device. The HI STAR has been designed to produce transverse, sagittal, coronal, and oblique section images of the head, body and extremities of the human body. Contrast in these images is produced primarily by T1, T2, and proton density tissue characteristics. This imager employs conventional MRI imaging techniques in which the application of a combination of pulsed radio frequency energy and magnetic gradient fields is applied to the body to generate MR proton echo signals. These multidimensionally coded signals are converted by Fourier transform into 2D and 3D image sets.

Here's an analysis of the provided text regarding acceptance criteria and the study that proves the device meets those criteria:

It's important to note that this document is a 510(k) Summary from 1997, specifically for a Magnetic Resonance Imager (MRI). The nature of acceptance criteria and supporting studies for such a device at that time, and compared to a more modern AI/software-as-a-medical-device (SaMD) submission, will be different. This document focuses on hardware performance and safety rather than diagnostic accuracy of an AI algorithm.

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

The document provides "Safety Parameters" and "Imaging Performance Parameters" which function as acceptance criteria for the MRI device.

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

The document states: "Sample images included with this submission were obtained on R &D systems at Health Sample Includio war this oubmisers. Each volunteer signed a consent form. There were no adverse reactions noted on any of the participants."

• Sample Size: The exact number of "volunteers" is not specified. It refers to "sample images" and "each volunteer," implying a limited, unnamed number.
• Data Provenance: The images were obtained on R&D systems at Health Images itself. It was collected from "volunteers" who signed consent forms, suggesting these were healthy volunteers rather than patients with specific pathologies. The study appears to be prospective in nature, as images were "obtained" for the submission, rather than retrospective analysis of existing data. The country of origin is implicitly the United States, given the submitter's address and FDA context.

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

This information is not provided in the document. For an MRI device submission of this nature (hardware performance), the "ground truth" for imaging parameters (SNR, uniformity, resolution, distortion) is typically established through quantitative measurements using phantoms and standardized protocols (e.g., NEMA standards), not expert interpretation of clinical images. The safety parameters are also measured quantitatively against established limits.

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

This is not applicable and therefore not provided. Adjudication methods are relevant for studies requiring multiple expert interpretations to establish a consensus ground truth for diagnostic AI, which is not the type of study described here.

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

This is not applicable and therefore not provided. This document pertains to the performance of an MRI machine itself, not an AI-assisted diagnostic tool. There is no mention of human readers or AI assistance in the context of diagnostic interpretation.

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

This is not applicable and therefore not provided. This document is for an MRI hardware device, not a standalone AI algorithm.

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

The "ground truth" for the performance metrics appears to be based on:

• Physical measurements against industry standards/specifications: For imaging parameters like SNR, image uniformity, slice thickness, resolution, and geometric distortion, the device's performance is compared against NEMA and AAPM standards. These involve quantitative measurements using phantoms, not clinical ground truth like pathology or expert consensus.
• Quantitative measurements against regulatory limits: For safety parameters such as magnetic field strength, SAR, and acoustic noise, physical measurements are taken and compared directly to FDA limits and IEC standards.

8. The sample size for the training set

This is not applicable and therefore not provided. This document describes an MRI hardware device, not a machine learning model that requires a training set.

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

This is not applicable and therefore not provided. As there is no training set for an AI algorithm mentioned, there is no ground truth established for one.

Ask a specific question about this device