Virtual Transcriptomics for Atherosclerotic Plaque Phenotyping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic methods for atherosclerosis lack non-invasive, patient-specific assessment of molecular features, limiting personalized pharmacotherapy and accurate identification of unstable atherosclerotic plaques, which are major contributors to cardiovascular disease.
Innovation Solution
Development of methods using machine intelligence models to interpret conventional imaging data from atherosclerotic plaques, combining virtual tissue modeling and transcriptomics to predict gene expression levels, providing patient-specific molecular phenotypes and plaque characterization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If non-invasive imaging methods are used to assess atherosclerotic plaques, then patient safety and comfort are improved, but molecular-level diagnostic precision is lost
Solution Approach 1:
The patent creates virtual copies of tissue transcriptomes through computational modeling. Machine learning models trained on paired imaging-transcriptome data generate virtual transcriptomic profiles from conventional imaging data, enabling molecular-level assessment without physical tissue sampling. This copying approach preserves the molecular information in virtual form while avoiding invasive biopsies.
Solution Approach 2:
The patent replaces mechanical tissue sampling (biopsy) with computational inference. Instead of physically extracting and analyzing tissue, the system uses machine learning models to infer transcriptomic profiles from imaging data. This substitution eliminates the need for invasive mechanical procedures while maintaining molecular-level diagnostic capability.
2Ease of operation
If population-based scoring methods are used for risk management, then diagnostic simplicity is maintained, but patient-specific accuracy is reduced
Solution Approach 1:
The patent transitions from population-level average assessments to patient-specific localized analysis. The machine learning models generate individualized virtual transcriptomes for each patient's plaque, capturing unique molecular characteristics. This local quality approach enables personalized risk stratification while maintaining computational feasibility through automated processing.
3Measurement precision
If tissue biopsies are obtained for molecular analysis, then transcriptomic accuracy is improved, but procedural complexity and patient burden increase
Solution Approach 1:
The patent introduces computational modeling as an intermediary between imaging and transcriptomics. Instead of directly obtaining tissue for analysis, the system uses machine learning models as intermediaries to translate imaging data into virtual transcriptomic profiles. This intermediary approach preserves transcriptomic accuracy while eliminating the need for complex biopsy procedures.
Data Source
AI summary
Atherosclerotic plaque phenotyping by image data analysis, e.g., using conventional computed tomography angiography (CTA) or other imaging modalities can elucidate the molecular signature of atherosclerotic lesions on a per-patient basis.