DLBCL Cell of Origin Classification via DNA Mutational Signatures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the cell of origin (COO) in diffuse large B Cell lymphoma (DLBCL) are either not feasible due to the need for RNA or high tumor purity, or they do not reliably recapitulate prognostic findings.
Innovation Solution
A DNA-based classification model, COODC, using the FoundationOne Heme platform, which requires only 20% tumor content and does not need RNA, to predict COO by analyzing genomic features such as alterations in specific genes and copy number changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If RNA-based classification methods (microarray, Nanostring) are used to determine COO subtype, then measurement precision is improved, but the method becomes infeasible due to requirements for high tumor purity (60%+) and RNA quality
Solution Approach 1:
The invention changes the input parameter from RNA expression data to DNA mutation data. Specifically, it uses mutations in genes such as EZH2, IGH:BCL2, MYD88, and CD79B, along with copy number alterations and mutational signatures, to classify COO subtypes. This parameter change allows the method to work with lower tumor purity samples where RNA-based methods fail.
Solution Approach 2:
The invention replaces the RNA-based molecular classification system with a DNA-based mutational classification system. By substituting the biological material (RNA vs DNA) and the analytical approach (expression profiling vs mutational analysis), the method achieves feasibility in clinical settings where RNA quality and tumor purity are insufficient for traditional methods.
2Ease of operation
If IHC-based algorithms are used to approximate COO, then ease of operation is improved, but reliability deteriorates as they do not reliably recapitulate prognostic findings
Solution Approach 1:
The invention replaces immunohistochemistry-based phenotypic classification with DNA-based mutational classification. By analyzing specific mutations (EZH2, IGH:BCL2 translocations for GCB; MYD88, CD79B for ABC) and mutational signatures, the method provides more reliable prognostic information while maintaining operational simplicity through targeted genomic testing.
Solution Approach 2:
The invention changes the classification parameters from protein expression levels detected by IHC to specific DNA mutations and copy number alterations. This parameter change enables more accurate identification of COO subtypes with proven prognostic value, as the mutational profile directly reflects the underlying biology of GCB and ABC subtypes.
3Ease of operation
If DNA-based classification is developed to work with lower tumor purity, then ease of operation is improved, but measurement precision may worsen compared to RNA-based gold standard
Solution Approach 1:
The invention uses specific DNA parameters (mutations in EZH2, IGH:BCL2, MYD88, CD79B; copy number alterations of chromosome 6p and 9p; T>A and T>G transversion frequencies) that are distinct between COO subtypes. These parameters provide robust classification signals that maintain high concordance with RNA-based methods even in low tumor purity samples.
Solution Approach 2:
The invention uses mutational signatures, particularly AID-related signatures, as intermediary markers that reflect the biological origin of DLBCL cells. These mutational patterns serve as reliable intermediaries to infer COO subtype from DNA data, achieving approximately 89% concordance with Nanostring assay results.
Data Source
Figure 1A~1B
Figure 2A~2B
Figure 2C
AI summary
Provided are methods of determining cell of origin (COO) for diffuse large B Cell lymphoma (DLBCL) using DNA for the analysis. The methods include identification of DLBCL COO as activated B Cell (ABC) and germinal center B Cell (GCB).