Breast Ultrasound Lesion Frame Exclusion for Precise GTC Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasound diagnostic apparatuses cannot accurately distinguish between perilobular and edematous stroma in the mammary gland region, leading to inaccurate assessment of the glandular tissue component (GTC) ratio, which is a risk factor for breast cancer, especially when a suspected lesion region is present.
Innovation Solution
The apparatus continuously acquires ultrasound images, detects suspected lesion regions, generates an evaluation target frame group excluding frames with detected lesions, and performs glandular tissue component evaluation on frames before and after the lesion detection, using machine learning and image analysis to differentiate between low- and high-echo regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If ultrasound images including suspected lesion regions are used for GTC evaluation, then the evaluation can be performed on available images, but the accuracy of GTC ratio measurement deteriorates because lesion regions interfere with stroma differentiation
Solution Approach 1:
The patent extracts and removes frames containing suspected lesion regions from the evaluation target frame group. The lesion detection unit identifies frames with lesions, and the target frame generation unit excludes these frames from GTC evaluation, ensuring that only clean frames without lesions are used for accurate stroma differentiation and GTC ratio measurement.
Solution Approach 2:
The patent performs preliminary lesion detection and frame selection before conducting GTC evaluation. By pre-identifying and excluding frames with suspected lesions, the system prepares a clean evaluation dataset in advance, preventing contamination of GTC measurements by lesion artifacts.
2Measurement precision
If machine learning and image analysis are used to differentiate stroma types, then the differentiation accuracy improves, but the device complexity increases
Solution Approach 1:
The patent replaces manual or simple threshold-based image analysis with machine learning-based automated analysis. The evaluation unit uses trained models to automatically differentiate perilobular stroma from edematous stroma based on ultrasound image features, achieving high differentiation accuracy without requiring complex manual intervention.
Solution Approach 2:
The system performs self-service through automated machine learning models that independently analyze ultrasound images, detect stroma types, and calculate GTC ratios without requiring external expert intervention for each measurement, thereby managing complexity internally while maintaining high precision.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate assessment of the GTC ratio and cancer risk in the mammary gland region, even with suspected lesions, by improving the differentiation between stroma types and providing precise glandular tissue component evaluation.
Implementation Method 1
an ultrasound beam is transmitted from the ultrasound probe toward a subject, an ultrasound echo from the subject is received by the ultrasound probe
Implementation Method 2
the perilobular stroma... includes many collagen fibers... the edematous stroma... is rich in extracellular matrix, with a mixture of collagen fibers and fat, and contains fewer collagen fibers
Data Source
Figure 1
Figure 2~3
Figure 4~6
AI summary
An ultrasound diagnostic apparatus includes: an image acquisition unit (33) that continuously acquires ultrasound images of a plurality of frames in which a mammary gland region of a subject is imaged; a lesion detection unit (26) that detects a suspected lesion region in the mammary gland region for each of the ultrasound images of the plurality of frames; a target frame generation unit (27) that generates an evaluation target frame group with ultrasound images of frames other than a frame in which the suspected lesion region is detected among the ultrasound images of the plurality of frames; and an evaluation unit (29) that performs a glandular tissue component evaluation on the ultrasound image of each frame in the evaluation target frame group.