Integrating molecular profiles into clinical frameworks through the Molecular Oncology Almanac to prospectively guide...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Current tumor molecular profiling focuses on single gene variants, limiting therapeutic strategy development.
• Interactions between genomic alterations and global molecular features ('second-order' alterations) impact clinical outcomes but are not integrated into interpretation algorithms.
• There is a need for advanced computational tools to interpret multimodal genomic data for precision oncology.

Purpose of the Study:

• To introduce the Molecular Oncology Almanac (MOAlmanac), an integrated algorithm and knowledge base for interpreting multimodal genomic data.
• To enable point-of-care clinical decision-making and generate translational research hypotheses in oncology.
• To provide an open-source computational method for individualized molecular profile interpretation.

Main Methods:

• Development of the MOAlmanac, a paired clinical interpretation algorithm and knowledge base.
• Benchmarking MOAlmanac against a first-order interpretation method using retrospective cohorts.
• Application of MOAlmanac to a prospective precision oncology trial cohort.

Main Results:

• MOAlmanac increased the number of clinical hypotheses generated compared to first-order methods.
• Evaluation of molecular features and profile-to-cell line matchmaking improved hypothesis generation.
• In a prospective trial, MOAlmanac nominated a median of two therapies per patient, with 47% receiving identified therapeutic strategies.

Conclusions:

• MOAlmanac provides an open-source computational method for integrative interpretation of individualized molecular profiles.
• The tool facilitates enhanced clinical decision-making and hypothesis generation in precision oncology.
• Integrative analysis of multimodal genomic data holds significant potential for advancing cancer treatment strategies.