Method and system for predicting efficacy of neoadjuvant chemotherapy for locally advanced gastric cancer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting the efficacy of neoadjuvant chemotherapy for locally advanced gastric cancer are inaccurate, with high rates of omission and misdetection, poor bio-interpretability, and failure to capture disease features effectively, leading to biased predictions and potential chemotherapy toxicity.
Innovation Solution
A method and system using non-Gaussian diffusion-weighted imaging with multi-b-values to calculate intravoxel incoherent motion and diffusion kurtosis models combined with a convolutional neural network and long short-term memory model for precise prediction of chemotherapy efficacy, involving image segmentation, data labeling, and deep perception networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radiomics machine learning models are used to predict chemotherapy response, then noninvasive identification of responsive patients is achieved, but prediction accuracy is insufficient with high omission and misdetection rates
Solution Approach 1:
The patent transforms conventional radiomics features into diffusion kurtosis model parameters (DKI) and intravoxel incoherent motion parameters (IVIM), which are more sensitive to tumor cellularity and heterogeneity. This parameter transformation enables the model to capture subtle changes in tumor tissue properties that conventional radiomics misses, thereby improving prediction accuracy and reducing omission/misdetection rates
Solution Approach 2:
The patent combines multiple imaging modalities (conventional CT radiomics features + diffusion kurtosis parameters + IVIM parameters) into a composite prediction model. This multi-modal integration allows the system to leverage the complementary strengths of each imaging technique, achieving more reliable and accurate predictions than any single modality could provide alone
2Loss of information
If conventional radiomics features are extracted from CT images, then tumor texture and spatial patterns are captured, but bio-interpretability of features is poor and disease sensitivity is insufficient
Solution Approach 1:
The patent replaces conventional radiomics feature extraction with diffusion physics-based models (DKI and IVIM) that directly model water molecule motion in tissue. This substitution provides both quantitative biological meaning (cellularity, heterogeneity, perfusion) and diagnostic accuracy, eliminating the black-box nature of conventional radiomics while maintaining noninvasive imaging capabilities
Solution Approach 2:
The patent transforms image intensity data into physically meaningful diffusion parameters (mean diffusion coefficient, diffusion kurtosis, perfusion fraction, pseudotransport coefficient). These parameters have direct biological interpretations related to tumor cellularity, heterogeneity, and blood flow, making the model both accurate and interpretable without requiring complex feature engineering
3Measurement precision
If multi-b-values non-Gaussian diffusion imaging is used, then cell proliferation activity and tumor heterogeneity are better reflected, but processing complexity and computational requirements increase
Solution Approach 1:
The patent segments the diffusion imaging process into distinct modeling stages: (1) acquiring multi-b-value DWI data, (2) fitting IVIM model to separate perfusion and diffusion components, (3) fitting DKI model to quantify non-Gaussian diffusion, and (4) extracting tumor regions using segmentation algorithms. This segmentation of the complex processing pipeline into modular steps makes the system more manageable and computationally efficient while maintaining high precision in heterogeneity detection
Solution Approach 2:
The patent introduces automated model fitting algorithms and machine learning classifiers as intermediaries between the raw diffusion imaging data and the final prediction. These intermediaries automatically handle the complex mathematical modeling and pattern recognition, reducing the computational burden on the user while maintaining the precision benefits of multi-b-values non-Gaussian diffusion imaging
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
Enhances the accuracy of predicting chemotherapy efficacy by automatically processing MRI data, extracting key tumor information, and providing a reliable diagnostic basis for clinicians, improving work efficiency and scientific research in medical imaging and prediction.
Implementation Method 1
non-Gaussian diffusion-weighted imaging with multi-b-values can obtain multiple quantitative parameters, such as intravoxel incoherent motion model and the diffusion kurtosis model, respectively, reflecting the cell proliferation activity, blood perfusion and tumor heterogeneity of the tumor tissue
Implementation Method 2
non-Gaussian diffusion-weighted imaging with multi-b-values can obtain multiple quantitative parameters, such as intravoxel incoherent motion model and the diffusion kurtosis model, respectively, reflecting the cell proliferation activity, blood perfusion and tumor heterogeneity of the tumor tissue
Data Source
AI summary
A method and a system for predicting the efficacy of neoadjuvant chemotherapy for locally advanced gastric cancer are provided. The method includes the following steps: the historical image data acquired through multi-b-values non-Gaussian diffusion MRI and corresponding prognostic image data are acquired, so that to obtain the signal intensity data corresponding to different b values, and influencing factors are acquired by combining diffusion models; a tumor tissue image is selected the region of interest which is further performed data labeling and enhancement to obtain a label set; the deep perception network is established to divide efficacy levels; A data set is established according to the influencing factors and the efficacy levels, a CNN-LSTM prediction model is constructed, the CNN-LSTM prediction model is optimized by using the data set to obtain an optimal model, and the chemotherapy efficacy level is evaluated by the optimal model.