(20 days)
The WorkstationOne™ Breast Imaging Workstation is intended for use with a regionally approved digital mammography system. The workstation displays images from multiple modalities, which include X-ray mammography MG, breast US and breast MRI. The workstation allows selection, display, manipulation, quantification, markup, print composition and media exchange of breast images. Here, quantification refers to measurements (such as area and distance) within a region of interest that the radiologist manually draws on the images. Similarly, markup refers to graphics that are manually drawn by the radiologist to indicate a region of interest. Note that the region of interest is not automatically generated by the computer.
The WorkstationOne™ Breast Imaging Workstation is intended for softcopy reading and interpretation of digital mammography images by Radiologists.
The WorkstationOne™ Breast Imaging Workstation when used for interpretation of images acquired using a FFDM image acquisition system shall display the images only on FDA-cleared high-resolution monitors. The images used for primary diagnostic reading must be in a lossless format, unless lossy formats are approved for use in digital mammography.
The WorkstationOne™ is a diagnostic breast imaging workstation which consists of a software system that obtains breast imaging screening or diagnosis exams from PACS or Modalities; displays and manipulates the images for radiologists to perform interpretation task. The workstation also supports media exchange and film printing. Optionally, the workstation can interface with a reporting system to generate an interpretation report.
The enterprise workflow of the workstation follows IHE integration profiles, specifically, MAMMO (Mammography Image Profile) and RWP (Reporting Workflow Profile). The workstation can be configured to use mammographic specific hanging protocol and reading workflow. The workstation obtains the source images and CAD reports either as the recipient of the push of that data, or by querying and retrieving them from a PACS archive. In both models, DICOM is used. The images can only be the lossless compressed or non-compressed DICOM images.
The WorkstationOne™ is a software system that can be installed on an off-the-shelf general-purpose computer with one or two aray-scale high-resolution monitors and one color monitor. The high-resolution monitors are used to display digital mammography images and associated overlays for the purpose of primary interpretation by radiologists. The color monitor is used for selecting studies, displaying color images, navigating workflow, and other user interface elements. Optionally, a dedicated keypad or touchpad is included for ergonomic reasons.
The provided 510(k) summary for the WorkstationOne™ Breast Imaging Workstation (K073272) does not include specific acceptance criteria or a detailed study proving the device meets acceptance criteria in the format typically associated with performance claims for AI/CAD devices. This submission predates the extensive regulatory guidance on clinical performance studies for AI/ML devices.
However, based on the information provided, we can infer the approach taken:
1. Acceptance Criteria and Reported Device Performance:
The submission explicitly states:
- "The software testing procedure has been developed. The procedure with pass/fail criteria has been run to ensure that the product meets all the specified requirements."
This indicates that internal acceptance criteria were established, likely focusing on technical specifications, functionality, and safety rather than a clinical performance metric like sensitivity or specificity. Given that it's a diagnostic workstation for displaying and manipulating images, the "performance" would be related to:
- Image Display Quality: Ensuring accurate and lossless display of various breast imaging modalities (MG, US, MRI) on FDA-cleared high-resolution monitors.
- Functionality: Correct operation of selection, manipulation, quantification, markup, print composition, media exchange, and integration with PACS/modalities/reporting systems.
- Safety: Adherence to a Risk Management Plan.
Since no specific quantitative performance metrics (e.g., sensitivity, specificity, AUC) are reported, a table comparing those to acceptance criteria cannot be generated as it would for an AI algorithm making a diagnostic claim. The "device performance" in this context refers to its ability to meet its functional and technical specifications as a workstation.
No quantitative performance criteria or results are provided in the document for comparison.
2. Sample Size for Test Set and Data Provenance:
The document describes software testing procedures, but does not specify a sample size for a "test set" of medical images or its provenance. The testing appears to be internal software validation rather than a clinical performance study using patient data.
3. Number of Experts and Qualifications for Ground Truth:
The document does not mention the use of experts to establish ground truth for a test set, nor does it specify their number or qualifications. This is likely because the device is a workstation for display and manipulation, not an AI algorithm making a diagnostic interpretation that requires expert ground truth for performance evaluation.
4. Adjudication Method for Test Set:
Since no test set with expert ground truth is described, no adjudication method is mentioned.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
No MRMC comparative effectiveness study is mentioned. The device is a workstation for interpretation, not an AI assisting human readers to improve their diagnostic accuracy. Its purpose is to provide the tools for interpretation.
6. Standalone (Algorithm Only) Performance Study:
No standalone performance study for an algorithm is mentioned. This device is a "Picture archiving and communications system" and a "diagnostic breast imaging workstation," meaning its primary function is to display images and provide tools for radiologists. It does not contain an AI algorithm that makes a diagnostic interpretation on its own. The mention of CAD reports being displayed indicates it can integrate with CAD systems, but it is not a CAD system itself.
7. Type of Ground Truth Used:
As established, the submission does not describe a study involving medical images requiring clinical ground truth (e.g., expert consensus, pathology, outcomes data). The "ground truth" for the software testing would have been against defined functional requirements and technical specifications.
8. Sample Size for Training Set:
No training set is mentioned or applicable as this device is not an AI/ML algorithm that learns from data.
9. How Ground Truth for Training Set Was Established:
Not applicable as there is no training set for an AI/ML algorithm.
Summary of Study (Based on 510(k) Text):
The "study" described in the 510(k) is an internal Non-clinical Test that involved:
- Risk Management Plan: Identifying and controlling potential hazards.
- Software Testing Procedure: Developing and executing a procedure with pass/fail criteria to ensure the product meets all specified requirements.
This type of testing focuses on:
- Functional Verification: Ensuring all features (selection, display, manipulation, quantification, markup, printing, media exchange, PACS/modality interface) work as intended.
- Technical Compliance: Adherence to standards like DICOM and IHE profiles.
- Safety and Reliability: Confirming the software's stability and absence of critical bugs.
The conclusion drawn from this "study" was that "The materials provided in this 510(k) submission have demonstrated the device is as safe, as effective, and performs as well as the predicate devices." This substantial equivalence argument is based on similar technological characteristics and the non-clinical testing performed to ensure its functionality and safety met internal criteria. It's crucial to understand that this is a 2007 submission for a workstation, not an AI diagnostic algorithm, hence the absence of clinical performance metrics and studies typical for AI devices today.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).