Metabolite Predictors of Breast and Colorectal Cancer Risk in the Women's Health Initiative

Area of Science:

• Metabolomics
• Biomarkers
• Cancer Epidemiology

Background:

• Metabolomics offers a window into the exposome, reflecting environmental exposures and metabolic processes.
• Understanding metabolic signatures may enhance early cancer detection and risk prediction.

Purpose of the Study:

• To evaluate the predictive power of prospectively measured metabolites for breast cancer (BC) and colorectal cancer (CRC).
• To determine if metabolites improve cancer risk prediction beyond established risk factors.

Main Methods:

• Analysis of serum and urine samples from 758 women with BC or CRC and 758 controls using LC-MS/MS, lipidomics, GC-MS, and NMR.
• Application of LASSO regression and Super Learner algorithms for feature selection and prediction model development.
• Performance assessment using cross-validation (CV) to determine predictive accuracy (CV-AUC).

Main Results:

• Metabolites did not enhance predictive performance for breast cancer (CV-AUC ~0.57).
• Metabolites significantly increased predictive performance for colorectal cancer (median CV-AUC increased from ~0.54 to ~0.60).
• Key metabolites contributing to CRC prediction included those related to energy, protein, and dietary/microbial metabolism.

Conclusions:

• Prospectively measured metabolites show potential for improving colorectal cancer risk prediction.
• Energy metabolism-related metabolites may play a crucial role in early colorectal cancer development.
• Metabolic profiling may offer insights into cancer development years before clinical diagnosis.