Urine cfDNA Sequencing for Bladder Cancer MRD Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting molecular residual disease (MRD) in bladder cancer patients are inefficient, leading to potential over- or under-treatment, as they lack accuracy in assessing residual disease post-treatment.
Innovation Solution
The method involves obtaining a urine sample, extracting cell-free DNA, and using ultra-low-pass whole genome sequencing (ULP-WGS) and next-generation sequencing (NGS) to detect tumor-derived metrics such as tumor fraction, variant allele frequency, and tumor mutational burden, combined with a machine learning model to determine the presence of residual disease.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current methods are used for detecting molecular residual disease, then the detection process is simple, but the measurement precision and reliability are insufficient leading to over- or under-treatment
Solution Approach 1:
The detection method is segmented into multiple specialized components: ultra-low-pass whole genome sequencing for comprehensive genomic coverage, targeted next-generation sequencing for specific tumor markers, and machine learning models for data interpretation. Each component addresses specific detection challenges, collectively achieving high precision without requiring a single overly complex system.
Solution Approach 2:
The method performs preliminary genomic profiling and establishes baseline tumor characteristics before treatment. This preliminary action enables the system to detect even minimal residual disease by comparing post-treatment samples against the established baseline, significantly improving detection accuracy while maintaining manageable complexity through pre-characterization.
2Reliability
If invasive surgeries like radical cystectomy are performed, then treatment thoroughness is high, but patient morbidity and quality of life deteriorate
Solution Approach 1:
The patent performs preliminary molecular assessment to identify patients with no residual disease before making treatment decisions. This preliminary action allows clinicians to confidently select less invasive treatments for patients who truly have no residual disease, while reserving radical surgery for those who need it, thereby improving treatment effectiveness while reducing unnecessary patient morbidity.
Solution Approach 2:
The high-sensitivity detection method serves as a self-selection tool, automatically identifying patients who require aggressive treatment versus those who can undergo conservative management. This self-service capability enables precise patient stratification without requiring complex clinical judgment, reliably matching treatment intensity to actual disease burden while minimizing harm to patients.
3Measurement precision
If high-sensitivity detection methods are implemented, then detection precision improves, but the complexity of the detection system increases
Solution Approach 1:
The complex detection system is segmented into modular components that can be independently optimized and validated: ultra-low-pass WGS for genomic landscape, targeted NGS for specific markers, and separate machine learning models for different analytical tasks. This segmentation allows each component to achieve high precision in its specific function while the overall system remains manageable through modular architecture.
Solution Approach 2:
Machine learning models serve as intermediaries between the raw sequencing data and clinical interpretation. These intermediaries automatically process complex genomic data, identify subtle patterns indicating residual disease, and translate technical measurements into clinically actionable results. This intermediary layer manages the complexity of high-sensitivity detection while preserving detection accuracy.
Data Source
AI summary
Low-pass whole genome sequencing, as well as targeted hybrid-capture next-generation sequencing, were performed to detect both small mutations and genome-wide copy number alterations to more precisely detect MRD after physician's-choice neoadjuvant chemotherapy. A machine learning model using both cell-free DNA and pre-treatment clinical characteristics is disclosed. With this method, it was possible to significantly predict both progression-free and overall survival.