CT Image Analysis Device Nodule Vessel Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lung cancer detection methods using low-dose helical CT scans face challenges in accurately distinguishing nodular abnormalities from lung blood vessels due to similar characteristics, leading to inefficiencies in diagnosis and increased burdens on medical professionals.
Innovation Solution
A computer-aided image diagnostic processing device and program that utilize a multi-slice CT system to specify nodule candidate regions, generate ellipsoidal models, and calculate a diminution index to assist medical doctors in determining whether a region is a nodule or a lung blood vessel, thereby improving diagnostic efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-detector helical CT is used to image the entire lung, then detection capability for small lung cancer is improved, but the number of images generated increases to several-hundreds, increasing the burden on diagnosing reading
Solution Approach 1:
The patent segments the large number of CT images into multiple groups based on anatomical position and imaging characteristics. By dividing the comprehensive image set into manageable segments, radiologists can systematically review each group without being overwhelmed by the total volume of several-hundreds of images, thus maintaining high detection capability while reducing reading burden.
Solution Approach 2:
The patent introduces a new dimension of organization by grouping images not just sequentially but by anatomical regions and diagnostic relevance. This multi-dimensional classification transforms the flat list of several-hundreds images into a structured hierarchy, enabling radiologists to navigate and assess findings more efficiently across the entire lung volume.
2Extent of automation
If characteristic amounts of nodule candidate region are used for discrimination, then automated detection is improved, but accuracy in distinguishing nodule from lung blood vessel deteriorates due to similar characteristics
Solution Approach 1:
The patent introduces an intermediary classification stage between automated nodule detection and final diagnosis. The system first identifies nodule candidate regions using automated detection, then applies additional discriminative features and classification rules to distinguish true nodules from blood vessels. This intermediary step resolves the ambiguity caused by similar characteristics while preserving automated detection efficiency.
Solution Approach 2:
The patent extends the feature set beyond basic nodule candidate characteristics by incorporating additional parameters such as vascular connectivity analysis, texture patterns, and spatial relationship features. By changing and expanding the parameter space used for discrimination, the system achieves higher accuracy in distinguishing nodules from blood vessels while maintaining automated detection capability.
3Ease of operation
If cross section image observation is used for final judgment, then simplicity of operation is improved, but time required for discrimination between nodule and lung blood vessel increases
Solution Approach 1:
The patent performs preliminary automated analysis and classification before presenting images to radiologists for final judgment. By pre-processing the images with automated detection and discrimination algorithms, the system prepares and prioritizes candidate regions in advance, allowing radiologists to focus their attention on pre-identified areas of interest rather than reviewing all images from scratch, thus reducing judgment time while maintaining operational simplicity.
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
A computer-aided image diagnostic processing apparatus (1) includes a storage unit (11) which stores a medical image representing the inside of a subject, a unit (11, 12, 13) which specifies an anatomical abnormality candidate region included in the medical image, and a generation unit (15) which generates a display image representing the abnormality candidate region and its peripheral region to be discriminable from each other based on the medical image.