Gene Expression Classifier for Venetoclax-Resistant AML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current treatments for acute myeloid leukemia (AML) using a combination of venetoclax and azacitidine are ineffective for approximately 30% of patients, necessitating a method to identify resistant subjects and provide alternative therapies.
Innovation Solution
Measuring the expression levels of specific genes in leukemia cells to classify resistance or responsiveness to venetoclax and azacitidine treatment, using machine learning classifiers to determine a score and compare it to a cutoff value for identifying resistant subjects, and recommending alternative therapies based on gene expression levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If venetoclax and azacitidine combination treatment is administered to AML patients, then complete remission is achieved in approximately 70% of patients, but approximately 30% of patients remain resistant and do not achieve remission
Solution Approach 1:
The patent applies preliminary action by measuring gene expression levels in leukemia cells before administering venetoclax and azacitidine treatment. This pre-treatment assessment identifies patients who are likely to be resistant, allowing clinicians to select alternative therapies before the standard treatment fails, thereby avoiding the 30% resistance problem proactively rather than reactively.
2Measurement precision
If gene expression levels of at least 10 specific genes are measured in leukemia cells, then accurate identification of resistant subjects is achieved, but treatment decision complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the complex task of predicting treatment resistance into measurable components - specifically, the expression levels of at least 10 discrete genes (including BCL2L11, BCL2, MCL1, and others). Each gene's expression level serves as an independent measurable parameter that can be quantified and combined to produce an overall resistance prediction, making the complex diagnostic task manageable and objective.
3Extent of automation
If machine learning classifiers are used to determine resistance scores based on gene expression levels, then automated and objective treatment classification is achieved, but computational requirements and analysis complexity increase
Solution Approach 1:
The patent applies the intermediary principle by introducing a machine learning classifier as a mediating computational system between the raw gene expression data and the clinical treatment decision. The classifier processes the complex multi-gene expression patterns and translates them into an interpretable resistance score or category, serving as an intermediary that automates the classification while presenting results in a clinically actionable format.
Data Source
AI summary
The present disclosure provides methods of treating venetoclax-resistant acute myeloid leukemia, including methods of identifying alternative treatment targets in specific subsets of patients who would otherwise be resistant to treatment with venetoclax and/or azacitidine.