Nodule Circularity Computation for Lung CAD
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer-assisted diagnosis (CAD) systems for lung cancer using low-dose helical CT scans face challenges in accurately detecting nodular abnormalities near the pleura due to limitations in quantifying the roundness of lesions, leading to false negatives and increased false positives.
Innovation Solution
A diagnostic imaging support processing apparatus and program that determine a nodular region, approximate its contour with a polygonal line, compute the degree of circularity using areas and reference points, and adjust for the border between tissue and nodule regions to provide a unified quantification of roundness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a low-dose helical CT scan is used for lung cancer examination, then the radiation exposure is reduced and lung cancer detection rate is improved, but the workload of image interpretation is increased due to hundreds of images generated per scan
Solution Approach 1:
The patent replaces manual image interpretation with an automated computer-based system that processes CT images. The system automatically detects nodular abnormalities, calculates circularity values, and generates diagnostic support information, substituting the mechanical process of manual review with an automated computational system.
Solution Approach 2:
The system enables self-service by automatically analyzing CT images without requiring manual intervention. The automated detection and circularity calculation processes allow the imaging system to serve itself in terms of initial analysis, reducing the burden on radiologists while maintaining high detection rates.
2Reliability
If manual detection of nodular abnormalities is performed, then false positives and false negatives occur due to limitations in quantifying roundness, but automated detection requires complex circularity calculation algorithms
Solution Approach 1:
The patent changes the parameter used for nodule characterization from subjective visual assessment to an objective quantitative parameter - the circularity value. By calculating the ratio of inscribed circle area to nodule area, the system transforms the detection problem into a parameter-based classification that improves reliability while managing complexity through mathematical formulation.
Solution Approach 2:
The patent introduces an intermediary computational step - the circularity calculation - that mediates between raw image data and diagnostic conclusions. This intermediary parameter serves as a bridge that quantifies the roundness characteristic, enabling more accurate detection while structuring the complexity into manageable computational steps.
3Measurement precision
If traditional circularity calculation methods are used for nodules near the pleura, then false negatives increase due to inaccurate roundness quantification, but adjusting for tissue-nodule borders increases computational complexity
Solution Approach 1:
The patent applies local quality by treating the border region between tissue and nodule differently from the nodule interior. The system identifies and adjusts for border characteristics specifically where tissue-nodule interfaces occur, applying localized processing to improve measurement precision in critical regions without unnecessarily complicating the entire calculation process.
Solution Approach 2:
The patent performs preliminary action by pre-identifying and adjusting for border effects before final circularity calculation. By anticipating and correcting for tissue-nodule interface artifacts in advance, the system improves measurement precision proactively, preventing false negatives before they occur rather than correcting them afterward.
Data Source
AI summary
A diagnostic imaging support processing apparatus includes a nodular region determination unit which determines a nodular region included in an image showing the inside of a subject, a polygonal line approximation processing unit which obtains a plurality of nodes constituting a polygonal line that approximates a contour of the nodular region, a reference position determination unit which determines a position of a reference point, and a circularity computation unit which computes the degree of circularity by using areas of a plurality of regions determined based on the plurality of nodes and the reference point.