Integrative Analysis of Metabolic Programs Associated with Cancer Risk and Outcomes in the PLCO Cohort

Principal Investigator

Name
Blake Rushing

Degrees
PhD

Institution
University of North Carolina at Chapel Hill

Position Title
Assistant Professor

Email
blake_rushing@unc.edu

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-2035

Initial CDAS Request Approval
Mar 26, 2026

Title
Integrative Analysis of Metabolic Programs Associated with Cancer Risk and Outcomes in the PLCO Cohort

Summary
Metabolic reprogramming is a hallmark of cancer development and progression, yet the systemic metabolic alterations that precede tumor formation remain poorly understood. Large prospective cohorts with available metabolomics data a unique opportunity to identify circulating metabolic signatures that reflect early biological processes involved in cancer initiation and progression. The Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) represents one of the most valuable resources for such analyses due to its large sample size, long-term follow-up, and extensive clinical, epidemiologic, and molecular data.

The goal of this project is to integrate metabolomics datasets available through PLCO with clinical, epidemiologic, genetic, and environmental exposure data to identify metabolic programs associated with cancer risk, progression, and outcomes. These analyses will contribute to the development of the Human Cancer Metabolome Atlas, an ongoing R01-funded effort in our lab to systematically characterize metabolic alterations across diverse cancer types and biological contexts.

First, we will identify circulating metabolic signatures associated with future cancer risk using metabolomics data measured in pre-diagnostic biospecimens. By comparing metabolite profiles between individuals who subsequently developed cancer and matched controls, we will identify metabolites and metabolic pathways associated with risk across multiple tumor types represented in PLCO, including prostate, lung, colorectal, ovarian, and other incident cancers captured during follow-up.

Second, we will integrate metabolomics data with germline genetic data and other molecular information available in PLCO to identify host genetic and biological factors influencing systemic metabolism and cancer susceptibility. These analyses will enable identification of metabolite quantitative trait loci (mQTLs) and will help clarify potential causal relationships between metabolic pathways and cancer risk.

Third, we will evaluate how metabolic profiles interact with lifestyle and environmental factors captured in PLCO, including diet, smoking history, body composition, medication use, and other exposures. These analyses will help elucidate how environmental and behavioral factors influence metabolic pathways associated with cancer development.

Finally, we will develop integrative models that combine metabolomics with clinical and epidemiologic data to identify coordinated metabolic programs associated with cancer incidence and outcomes. Advanced statistical and machine learning approaches will be used to identify latent metabolic factors and pathway-level signatures that capture coordinated metabolic changes associated with cancer risk and progression.

All analyses will be conducted using de-identified data accessed through the NCI Cancer Data Access System in accordance with PLCO data use policies. Findings from this work will contribute to the goals of the Human Cancer Metabolome Atlas and will help advance understanding of systemic metabolic processes that influence cancer development, ultimately supporting improved strategies for cancer prevention, early detection, and precision oncology.

Aims

Cancer metabolic reprogramming is a recognized hallmark of tumor development, yet the systemic metabolic alterations that precede and drive cancer initiation remain poorly characterized. Large prospective cohorts with pre-diagnostic biospecimens and metabolomics data offer a rare opportunity to capture early circulating metabolic signals associated with cancer risk and outcomes. The Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) - with its large sample size, long-term follow-up, and rich clinical, epidemiologic, and molecular data - represents an ideal resource for such analyses. The overarching goal of this project is to integrate PLCO metabolomics data with genetic, clinical, epidemiologic, and environmental exposure data to identify metabolic programs associated with cancer risk, progression, and outcomes, contributing to the Human Cancer Metabolome Atlas, an R01-funded effort to systematically characterize metabolic alterations across cancer types.
• Aim 1: Identify pre-diagnostic circulating metabolic signatures associated with cancer risk. Using metabolomics data measured in pre-diagnostic biospecimens, we will compare metabolite profiles between individuals who subsequently developed cancer and matched controls to identify metabolites and metabolic pathways associated with risk across multiple tumor types represented in PLCO, including prostate, lung, colorectal, ovarian, and other incident cancers captured during follow-up.
• Aim 2: Integrate metabolomics with germline genetic data to identify host factors influencing metabolism and cancer susceptibility. We will integrate PLCO metabolomics data with available germline genetic and molecular data to identify metabolite quantitative trait loci (mQTLs) and clarify potential causal relationships between metabolic pathways and cancer risk, elucidating how host genetic factors shape systemic metabolism in the context of cancer development.
• Aim 3: Evaluate interactions between metabolic profiles and lifestyle and environmental exposures. We will assess how metabolic signatures interact with behavioral and environmental factors captured in PLCO - including diet, smoking history, body composition, and medication use - to elucidate how modifiable exposures influence cancer-associated metabolic pathways and inform potential prevention strategies.
• Aim 4: Develop integrative models combining metabolomics with clinical and epidemiologic data to characterize coordinated metabolic programs. Using advanced statistical and machine learning approaches, we will identify latent metabolic factors and pathway-level signatures that capture coordinated metabolic changes associated with cancer incidence and outcomes, generating multi-modal predictive models to support early detection and precision oncology efforts.

Collaborators

Blake Rushing University of North Carolina at Chapel Hill
Anh Moss University of North Carolina at Chapel Hill
Mengxuan Lai University of North Carolina at Chapel Hill
Yuheng Che University of North Carolina at Chapel Hill