CAD Mark Navigation for Tomosynthesis Mammography Review
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The diagnostic review of tomosynthesis mammography images is inefficient due to the need to manually navigate through multiple 2D images, which increases the time required for analysis and comparison with previous images, highlighting the need for improved navigation tools to enhance speed and accuracy.
Innovation Solution
The development of interactive visualization systems that utilize CAD marks to facilitate navigation between views of medical imaging data sets, allowing users to select and navigate to suspected abnormalities through a graphical user interface, with features like vicinity marks and automated viewing protocols to streamline the review process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a stack of 2D images is used to represent 3D breast tissue in tomosynthesis, then the 3D representation is achieved, but the review time is multiplied by the number of images in the stack
Solution Approach 1:
The patent extracts and highlights only the relevant slices containing CAD marks from the complete image stack, separating the diagnostically important information from the redundant images. This allows radiologists to focus on specific slices with suspected abnormalities rather than reviewing the entire stack sequentially.
Solution Approach 2:
CAD algorithms perform preliminary analysis of the image stack before radiologist review, automatically identifying and marking suspicious areas. This preliminary detection and annotation process prepares the data by highlighting regions of interest, so that when radiologists review the images, they can navigate directly to pre-identified areas of concern rather than examining each slice from scratch.
2Reliability
If manual navigation through multiple 2D images is performed, then complete review of the stack is possible, but the speed of review process decreases
Solution Approach 1:
The system provides feedback to radiologists by displaying CAD-generated marks and annotations on relevant slices, indicating where abnormalities have been detected. This feedback mechanism guides the radiologist's attention to specific areas, enabling faster navigation through the stack while maintaining review completeness, as the CAD system continuously informs the radiologist about suspicious regions that require examination.
Solution Approach 2:
The review interface dynamically adapts to radiologist needs by automatically navigating to and highlighting slices containing CAD marks. The system can dynamically adjust the display to show multiple relevant slices simultaneously or sequentially based on the detected abnormalities, making the review process more flexible and efficient compared to static manual navigation through all slices.
3Measurement precision
If CAD marks are used to identify suspected abnormalities, then detection accuracy is improved, but navigation complexity increases
Solution Approach 1:
The patent segments the image stack into discrete groups based on CAD marks, with each group representing a specific abnormality or region of interest. The navigation interface is segmented to display only the relevant slices containing marks, rather than presenting the entire stack. This segmentation reduces navigation complexity by organizing the large number of images into manageable, purpose-driven groups that radiologists can easily navigate through.
4Ease of operation
If comparison of right and left breast images is performed using mirrored setup, then visual comparison is simplified, but the ability to navigate to specific abnormalities across views is limited
Solution Approach 1:
The navigation system is designed to be universal, working with both traditional mirrored breast comparison views and CAD-marked slice navigation. It can display CAD marks across multiple views simultaneously, allowing radiologists to compare breasts using the mirrored setup while also navigating to specific abnormalities identified by CAD. The system integrates multiple viewing and navigation modes into a single flexible interface that adapts to different diagnostic workflows.
Data Source
AI summary
Systems, processors and methods display image views of stacks or volume image data and allow a user to navigate the data and/or views using CAD findings. The systems, methods or processors can generate vicinity CAD marks that have an appearance that varies depending (a) the proximity to the suspected abnormality and/or (b) the navigation direction required to move to the suspected abnormality. The visual appearance of a respective vicinity CAD mark in a view can correspond to proximity of the related suspected abnormality to a current view, whereby the vicinity CAD marks have higher intensity, color saturation, size, or other more dominant visual characteristic in views close to the suspected abnormality and lower intensity, color saturation, smaller size or other less dominant visual characteristic in views further away from the suspected abnormality.