Lung Nodule Segmentation and Registration for Growth Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting lung cancer, particularly in early stages, face challenges in distinguishing true pulmonary nodules from shadows, vessels, and ribs in low-dose computed tomography (LDCT) chest scans, and inadequately account for deformations due to heartbeat and respiration, leading to inaccurate segmentation and registration of lung tissues.
Innovation Solution
A computer-aided diagnostic system that employs a Linear Combination of Discrete Gaussians (LCDG) model and Markov Gibbs Random Field (MGRF) for improved lung segmentation, combined with global and local registration techniques to accurately align and measure volumetric changes in pulmonary nodules over time, compensating for positional and respiratory variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional nodule filtering or template matching is used to detect pulmonary nodules, then initial candidate selection can be performed, but false positive nodules cannot be reliably distinguished from true nodules due to overlapping shadows, vessels and ribs
Solution Approach 1:
The patent applies segmentation by separating the lung field into multiple regions based on anatomical structures. Lung regions are segmented to identify and exclude areas containing ribs, vessels, and shadows, thereby isolating true pulmonary nodules from false positive candidates. This regional segmentation approach directly addresses the difficulty of distinguishing nodules from overlapping structures.
2Measurement precision
If chest X-ray screening is used to detect early lung cancer, then large malignant nodules can be detected, but many false-positive test results are produced causing needless extra tests
Solution Approach 1:
The patent applies local quality by assigning different evaluation criteria and processing methods to different regions within the lung field. Each lung region is characterized by its specific anatomical features (e.g., proximity to ribs, vessels, or airways), and nodule candidates are evaluated according to their local regional characteristics. This allows for more accurate differentiation between true nodules and false positives in different anatomical contexts.
3Difficulty of detecting and measuring
If spherical, cylindrical or circular filters are used to detect lung nodules, then small lung nodules can be detected from high resolution CT images, but the general geometry of irregular lesions cannot be adequately described
Solution Approach 1:
The patent applies dynamics by using adaptive filtering kernels that change their shape and size based on the local characteristics of the image data. Rather than using fixed spherical or cylindrical filters, the system dynamically adjusts filter parameters to match the actual geometry of detected structures. This allows accurate detection and characterization of nodules with irregular shapes, spiculation, or varying orientations.
4Extent of automation
If morphological operators are used to detect lung nodules, then pattern recognition can be performed, but difficulties in detecting lung wall nodules remain
Solution Approach 1:
The patent applies dimensionality change by transitioning from 2D slice-based analysis to 3D volumetric analysis of lung regions. By processing the entire 3D lung volume and examining nodules from multiple spatial dimensions, the system can accurately detect and characterize lung wall nodules that may be missed in 2D cross-sections. This multi-dimensional approach provides better context for distinguishing true nodules from artifacts near the lung periphery.
Data Source
AI summary
A computer aided diagnostic system and automated method diagnose lung cancer through tracking of the growth rate of detected pulmonary nodules over time. The growth rate between first and second chest scans taken at first and second times is determined by segmenting out a nodule from its surrounding lung tissue and calculating the volume of the nodule only after the image data for lung tissue (which also includes image data for a nodule) has been segmented from the chest scans and the segmented lung tissue from the chest scans has been globally and locally aligned to compensate for positional variations in the chest scans as well as variations due to heartbeat and respiration during scanning. Segmentation may be performed using a segmentation technique that utilizes both intensity (color or grayscale) and spatial information, while registration may be performed using a registration technique that registers lung tissue represented in first and second data sets using both a global registration and a local registration to account for changes in a patient's orientation due in part to positional variances and variances due to heartbeat and/or respiration.