Radiomic Feature Analysis for Non-Invasive Immunotherapy Response Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting patient response to nivolumab immunotherapy for non-small cell lung cancer (NSCLC) are invasive, expensive, and have low accuracy, with conventional tissue-based markers being ineffective.
Innovation Solution
The use of computerized analysis of radiomic features extracted from computed tomography (CT) imagery to predict response to nivolumab or other checkpoint inhibitor immunotherapy in NSCLC patients, focusing on texture, intensity, and shape features in the tumoral and peritumoral regions, which are not visible to the human eye.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional tissue-based markers are used to predict patient response to nivolumab, then the assessment can be performed, but the accuracy is low and the method is ineffective
Solution Approach 1:
The patent replaces conventional tissue-based marker assessment with radiomic feature analysis from CT images. This substitution transitions from invasive tissue sampling to non-invasive imaging-based prediction, using computational extraction of texture, intensity, and shape features to achieve higher prediction accuracy while maintaining clinical effectiveness
Solution Approach 2:
The patent changes the parameters used for prediction from conventional tissue markers to radiomic features including texture features (e.g., entropy, homogeneity), intensity features (e.g., mean intensity, standard deviation), and shape features (e.g., sphericity, surface area). These parameter changes enable more accurate prediction of patient response to nivolumab
2Measurement precision
If invasive biopsy methods are used for prediction, then tissue samples can be obtained, but the procedure is invasive and expensive
Solution Approach 1:
The patent creates a virtual copy of the tumor characteristics by extracting radiomic features from CT images. Instead of physically obtaining tissue samples through biopsy, the system copies the relevant diagnostic information into quantitative features (texture, intensity, shape) that can be analyzed computationally, eliminating the need for invasive procedures while maintaining prediction capability
Solution Approach 2:
The patent introduces radiomic feature analysis as an intermediary between CT imaging and treatment response prediction. This intermediary layer extracts meaningful quantitative features from images, serving as a mediator that provides accurate prediction information without requiring direct tissue sampling, thereby reducing invasiveness and cost
3Measurement precision
If radiomic feature analysis is used, then non-invasive prediction with high accuracy is achieved, but the device complexity increases
Solution Approach 1:
The patent makes the CT imaging system multi-functional by enabling it to serve both its primary diagnostic purpose and the additional function of radiomic feature extraction for treatment response prediction. The same imaging data is reused to generate multiple types of information (anatomical visualization and quantitative prediction features), reducing the need for separate specialized equipment and minimizing additional system complexity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods, apparatus, and other embodiments predict response to immunotherapy from computed tomography (CT) images of a region of tissue demonstrating non-small cell lung cancer (NSCLC). One example apparatus includes a set of circuits that includes an image acquisition circuit that accesses a CT image of a region of tissue demonstrating cancerous pathology, a tumoral definition circuit that generates a tumoral surface boundary that defines a tumoral volume, a peritumoral segmentation circuit that generates a peritumoral region based on the tumoral surface boundary, and that segments the peritumoral region into a plurality of annular bands, a radiomics circuit that extracts a set of discriminative features from the tumoral volume and at least one of the plurality of annular bands, and a classification circuit that classifies the ROI as a responder or a non-responder, based, at least in part, on the set of discriminative features.