Search Results

The ResolutionMD™ Mobile software is a software-based Picture Archiving and Communication System (PACS) used with general purpose computing servers and specific mobile devices. It provides for communication, storage, reformatting, rendering on the server component and communication and display of DICOM 3.0-compliant CT and MR medical images as well as reports on the mobile device.

The ResolutionMD Mobile provides wireless and portable access to medical images. The device is intended for use as a diagnostic, review, and analysis tool by trained professionals such as radiologists, physicians and technologists. This device is not intended to replace full workstations and should be used only when there is no access to a workstation.

The ResolutionMD Mobile is not to be used for mammography.

The ResolutionMD™ Mobile 3.1 software is a software-based Picture Archiving and Communication System (PACS) used with general purpose computing servers and highresolution Apple Inc. iOS and Google Inc. Android OS-based wireless mobile devices for the display and advanced visualization of medical image data. It provides for communication, storage, processing, rendering on the server and the display of DICOM 3.0 compliant image data derived from CT and MRI on the mobile device.

The ResolutionMD Mobile 3.1 is a software-based Picture Archiving and Communication System (PACS) intended for use as a diagnostic, review, and analysis tool for CT and MR medical images on specific mobile devices. It is not intended to replace full workstations and should not be used for mammography.

Here's an analysis of the acceptance criteria and the study that proves the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance

The FDA submission for ResolutionMD Mobile 3.1 (K123186) primarily focuses on establishing substantial equivalence to a predicate device (ResolutionMD Mobile, K111346). The acceptance criteria for the new version, specifically supporting Android devices, revolve around demonstrating that the image quality and diagnostic confidence achieved on Android platforms are comparable to the previously cleared iOS platforms (predicate device) and adequate for clinical use.

• The Android mobile devices (smartphone and tablet) were comparable to the predicate iPhone and iPad devices.
• The devices were of adequate quality for clinical use.
• They were comfortable with the diagnoses made on the Android mobile devices using the ResolutionMD Mobile software.
• The overall clinical image display quality on the Android devices was equivalent to the iOS devices for the identification of clinically-relevant pathology.
• Comments on image contrast and sharpness included "very comparable" and "is diagnostic."
• No image artifacts were noted by the reviewers.
• The software and devices provide acceptable quality for regular use, and they were comfortable reviewing images. |

Study Proving Acceptance Criteria:

The study combined Performance Testing and Clinical Testing to demonstrate the device meets the acceptance criteria.

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

• Performance Testing Test Set: The sample size for the display performance tests was two mobile devices: an Android smartphone and an Android tablet.
  • Data Provenance: The tests were conducted by an ISO 17025-certified third party, implying controlled laboratory conditions and objective measurements. The data is "prospective" in the sense that the tests were specifically designed and executed to evaluate the new device's performance. The country of origin of the data is not explicitly stated, but the ISO certification and NIST traceability suggest an internationally recognized standard of testing.
• Clinical Testing Test Set: Not explicitly stated, but it involved a "series of typical CT and MR cases."
  • Data Provenance: The cases were reviewed by a panel of radiologists in the United States. The specific nature (retrospective/prospective) of the "series of typical CT and MR cases" is not detailed, but it likely involved retrospective cases from clinical archives for evaluation.

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

• Performance Testing: Ground truth was established by adherence to the AAPM Assessment of Display Performance for Medical Imaging Devices (2005) document. The "experts" in this context are the authors and contributors of the AAPM document, representing a consensus of medical physics and imaging professionals on display performance standards. The testing was carried out by an ISO 17025-certified third party, ensuring expertise in calibration and measurement.
• Clinical Testing: Three board-certified radiologists in the United States. Their specific years of experience are not mentioned, but "board-certified" indicates a recognized high level of expertise in diagnostic radiology.

4. Adjudication Method for the Test Set

• Performance Testing: The adjudication method was based on whether the devices "passed all of the tests" as defined by the AAPM document. This implies a pass/fail threshold for each of the nine display performance tests.
• Clinical Testing: The adjudication method was based on consensus among all three radiologists. The document repeatedly states "All three radiologists agreed that..." and refers to their collective findings. This indicates a form of unanimous consensus. There's no mention of a 2+1 or 3+1 rule; rather, it appears all three came to the same conclusion directly.

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 in the context of AI assistance. This submission is for a PACS visualization device, not an AI diagnostic algorithm. The clinical testing was a comparative assessment between human readers using the Android mobile devices vs. human readers using the predicate iOS devices to establish equivalence in image display quality and diagnostic confidence. It did not involve comparing human readers with and without AI assistance to measure an effect size related to AI improvement. The device itself is a display and analysis tool, not an AI-powered diagnostic aid.

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

• No, a standalone (algorithm-only) performance was not done. This device is a software-based PACS for image display and review by trained professionals. Its intended use inherently involves a "human-in-the-loop" for interpretation and diagnosis. The performance testing focused on the display capabilities of the device, and the clinical testing assessed human readers' diagnostic confidence and image quality perception when using the device.

7. The Type of Ground Truth Used

• Performance Testing: The ground truth was based on established technical standards and objective measurements as outlined in the AAPM Assessment of Display Performance for Medical Imaging Devices (2005). The test equipment and calibration were certified traceable to NIST, further solidifying the objectivity of the ground truth.
• Clinical Testing: The ground truth for comparative image quality and diagnostic confidence was expert consensus among the three board-certified radiologists comparing the Android display to the predicate iOS display for "clinically-relevant pathology." While the "clinically relevant pathology" implies there was a true pathological state in the cases, the primary "ground truth" for the equivalence assessment was the radiologists' agreement on the adequacy of the display for identifying that pathology.

8. The Sample Size for the Training Set

• Not Applicable / Not Provided. This submission primarily addresses the substantial equivalence of a new version of a PACS viewing software on Android devices to a previously cleared version on iOS. It describes the software's design and testing relative to established display standards and clinical usability. There is no mention of machine learning or AI components that would require a distinct "training set" in the traditional sense. The software's development would involve standard software engineering practices, verification, and validation, rather than a machine learning training phase.

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

• Not Applicable / Not Provided. As there's no mention of a distinct "training set" for a machine learning algorithm, there's no information on how a ground truth for such a set was established.

Ask a specific question about this device