Search Results

eRAD PACS is a software-only medical device used to receive medical images, scheduling information and clinical reports, organize and store them in an internal format, and make the information available across a network via web and customized user interfaces.

eRAD PACS includes software intended for use by qualified professionals for the presentation, review and comparison of diagnostic medical images.

eRAD PACS is for hospitals, imaging centers, reading practices, radiologists, technicians, physicians and other users who require and are granted access to patient image, exam and report information.

The eRAD PACS viewer displays images from DICOM compliant modalities and other devices including CT, computed and digital radiography, MRI, mammography, nuclear medicine, PET, secondary capture, ultrasound, x-ray angiography, x-ray fluoroscopy and visible light systems.

Lossy compressed images and digitized film images must not be used for primary diagnosis of mammography studies. When displaying mammography images for clinical interpretation, only monitors having regulatory clearance for mammography interpretation should be used.

eRAD PACS is a software-only medical image management and processing system, comprised of a central systems manager component, diagnostic viewing components, and an archiving component.

The system is used for patients who undergo an imaging procedure deemed necessary by the patient's physician.

The data flow is as follows:

• Patient and procedure information is optionally acquired by the central system manager to prepare for the acquisition of image objects.
• -Image objects are acquired from the image sources, such as imaging modalities, PACS, data archives, and other devices.
• After receiving the procedure information or the image objects, the central system manager searches for and retrieves relevant prior procedure data from the archiving component.
• When the central system manager registers the acquired image objects and the retrieved prior procedure data, a user can access the information from a workstation by selecting the item from the operator's worklist.
• The image data is transmitted to and rendered on the user's workstation using the diagnostic viewing components.
• -After reviewing the images in the diagnostic viewer, the user optionally creates a clinical report using a text editor or a commercially available speech recognition solution.
• -Once the central system manager registers a report, the report is available for access by the referring physician, or it can be exported into a third-party information system.
• At some configured point in time, the image data and the report information are delivered to the archiving component for backup and long-term storage.

The provided document, a 510(k) premarket notification for eRAD PACS, does not contain explicit acceptance criteria or a detailed study proving the device meets these criteria in the way a clinical performance study for an AI/ML medical device would.

The document primarily focuses on demonstrating substantial equivalence to a previously cleared predicate device (eRAD PACS, K120995). The crucial statement regarding testing is:

"Thorough non-clinical system verification and validation testing was conducted in accordance with applicable standards and internal design procedures to verify that the eRAD PACS software product meets user needs and its intended use. Testing demonstrated that the eRAD PACS software product is substantially equivalent to the predicate device."

This indicates that the "study" conducted was primarily non-clinical verification and validation (V&V) testing for a software-only medical image management and processing system (PACS), not a clinical performance study involving human readers or standalone algorithm performance against a clinical ground truth.

Therefore, for the specific questions asked, a direct answer cannot be fully provided from the given text as the nature of the submission (510(k) for a PACS update) does not require the same type of clinical performance data as, for example, an AI algorithm for disease detection.

However, I can deduce and infer information based on the context of a PACS 510(k) submission:

Analysis based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance:

• Acceptance Criteria: The document does not explicitly state quantitative acceptance criteria in terms of performance metrics (e.g., sensitivity, specificity, AUC) for the eRAD PACS. Instead, the acceptance criterion for this 510(k) submission is implicitly "substantial equivalence" to the predicate device for its intended use, as verified by non-clinical V&V testing. This means the device must function as intended, handle images correctly, and meet the relevant technical specifications and standards (e.g., DICOM compliance).
• Reported Device Performance: The document states: "Testing demonstrated that the eRAD PACS software product is substantially equivalent to the predicate device." This is the reported performance. Specific numerical metrics are not provided because the "performance" here refers to the system's ability to manage and display images functionally as a PACS, not to diagnostic accuracy.

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

• The document mentions "non-clinical system verification and validation testing." For a PACS system, this typically involves testing with a diverse set of synthetic and real-world DICOM images to ensure proper handling, display, and archiving across various modalities.
• Sample Size: Not specified. It would likely involve a large variety of DICOM images and system configurations to test different functionalities and edge cases.
• Data Provenance: Not specified. Given it's V&V for a PACS, the "data" would be medical images (DICOM files) from various modalities. It is likely a mix of internally generated test data, publicly available datasets, and potentially de-identified clinical data, but this is not stated. The provenance (country, retrospective/prospective) of these images is not detailed.

3. Number of Experts Used to Establish Ground Truth and Qualifications:

• Not applicable in the context of this 510(k) submission. Ground truth established by experts is typically for validation of diagnostic accuracy (e.g., for an AI algorithm interpreting images), not for the functional performance of a PACS.
• For PACS V&V, the "ground truth" is that the system correctly displays the image, stores it accurately, and performs its functions as specified. This is verified by engineers and testers against technical specifications, not by clinical experts establishing diagnostic "ground truth."

4. Adjudication Method for the Test Set:

• Not applicable. Adjudication (e.g., 2+1, 3+1) is a method used in clinical studies to establish a rigorous "ground truth" for diagnostic tasks, usually when there is observer variability. This is not mentioned as part of the PACS V&V.

5. MRMC Comparative Effectiveness Study:

• No evidence. The document does not describe an MRMC study comparing human readers with and without AI assistance. This type of study is relevant for AI algorithms intended to aid diagnosis, not for a PACS system whose primary function is image management and display.

6. Standalone Performance (Algorithm Only without Human-in-the-Loop):

• Not applicable. The eRAD PACS is described as a "medical image management and processing system" and a "viewer" for qualified professionals. It is not an AI algorithm performing a diagnostic task independently. Its "performance" is in its ability to correctly manage and display images.

7. Type of Ground Truth Used:

• Technical Specifications and DICOM Standards: For a PACS system, the "ground truth" for verification and validation is primarily defined by its technical specifications (e.g., image fidelity, display characteristics, network communication standards like DICOM, storage integrity) and user requirements. It's about whether the system functions correctly as an imaging management system, not about establishing clinical diagnostic truth from patient outcomes or pathology.

8. Sample Size for the Training Set:

• Not applicable. The eRAD PACS described is not an AI/ML device that requires a "training set" in the machine learning sense. It's a software system built based on established programming principles and standards.

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

• Not applicable. As a non-AI/ML PACS, there is no "training set" or corresponding ground truth establishment process in the way an AI model would have.

Summary based on context:

The eRAD PACS 510(k) submission describes an update to an existing PACS system, specifically a "restructure of the internal components for deployment in a cloud environment." For such a device, the regulatory burden focuses on ensuring that the changes do not undermine the safety and effectiveness established by the predicate device. This is achieved through non-clinical verification and validation testing against technical specifications and performance requirements of an image management system. It is not a clinical study involving diagnostic accuracy metrics or human reader performance.

Ask a specific question about this device