Genomic-driven nutritional interventions for radiotherapy-resistant rectal cancer patient

Area of Science:

• Oncology
• Nutritional Science
• Bioinformatics

Background:

• Radiotherapy response in rectal cancer is complex, involving stress response, cell death, and metabolism.
• Lipid metabolism modulation is a promising strategy to enhance radiotherapy outcomes, accessible via dietary bioactive molecules.
• Experimental identification of all potential radioresponse modulators is infeasible.

Purpose of the Study:

• To develop a machine learning (ML) approach to identify bioactive food molecules that modulate radiotherapy response.
• To design phytochemically-enriched recipes to improve radiotherapy outcomes for rectal cancer patients.
• To establish a pipeline for genomic-driven nutritional interventions.

Main Methods:

• Genomic-driven network ML, metric learning, and domain knowledge were used to identify potential radioresponse modulators.
• The Recipe1M database was utilized for recipe optimization.
• Ingredient substitutions were made to maximize predicted modulators while preserving culinary qualities.

Main Results:

• Identification of potential radioresponse modulators from bioactive food molecules.
• Development of optimized recipes incorporating these modulators.
• Demonstration of a computational approach for designing nutritional interventions.

Conclusions:

• Machine learning can effectively identify food-derived compounds to enhance radiotherapy response.
• Phytochemically-enriched recipes offer a novel strategy for improving rectal cancer patient outcomes.
• This work enables the design of targeted nutritional interventions for cancer therapy.