Single blood test could predict colorectal cancer recurrence, metastasis risk, researchers say

A single blood test may soon be enough to predict with greater accuracy whether colorectal cancer will recur or spread in a patient.

KAIST announced Wednesday that a research team led by Professor Lee Ji-min of the Graduate School of Medical Science and Engineering and Professor Kim Hyun-woo of the Department of Chemistry has developed a new framework for analyzing serum circulating amino acid networks — amino acids that travel through the bloodstream and reflect the body's metabolic state — in collaboration with Gangnam Severance Hospital and Asan Medical Center.

Using the framework, the team found that as colorectal cancer advances, the body's overall metabolic system undergoes a progressive reorganization, and proposed a new analytical strategy for predicting a patient's risk of recurrence and metastasis.

South Korea has the world's highest colorectal cancer incidence rate. A global colorectal cancer survey of 184 countries by the International Agency for Research on Cancer, a body under the WHO, found that South Koreans develop the disease at a rate of 45 cases per 100,000 people — the highest among all countries surveyed.

Amino acids are essential to cancer cell survival and proliferation, playing roles in energy production and DNA synthesis. Colorectal cancer is particularly known as a cancer in which amino acid metabolism changes significantly.

These changes are not confined to tumor tissue but also appear in the bloodstream, making circulating amino acids an important metabolic biomarker reflecting the body's overall metabolic condition.

Until now, however, research had focused on measuring how the concentration of individual amino acids changes, leaving largely unexplored how amino acids are interconnected and how they shift together.

The team applied 19F NMR (fluorine nuclear magnetic resonance spectroscopy) technology, which can simultaneously and precisely analyze 18 types of circulating amino acids from a small blood serum sample. By examining not only the concentration of each amino acid but also the relationships between them in network form, the team confirmed that the amino acid network in the blood undergoes a stepwise reorganization as colorectal cancer progresses.

As the cancer advanced, the proportion of branched-chain amino acids such as valine and leucine — which play key roles in muscle function and energy metabolism — declined, while the proportions of glycine and serine, which cancer cells rely on to synthesize DNA and proliferate rapidly, tended to increase.

The team then applied connectivity data from the amino acid network analysis to a machine learning model to assess recurrence and metastasis risk in colorectal cancer patients.

The method predicted recurrence and metastasis risk more accurately than carcinoembryonic antigen, the blood test marker widely used in hospitals to monitor colorectal cancer patients' progress. It also outperformed conventional methods that analyze only the individual concentrations of amino acids in the blood.

"We hope this will lead to a new precision medicine technology that can more accurately predict recurrence risk in colorectal cancer patients through a simple blood test and help establish personalized treatment strategies," Lee said.

Lee added that the long-term goal is to develop the approach into a blood-based supplementary diagnostic tool integrated with routine blood tests, minimizing the burden of additional blood draws on patients while supporting assessments of recurrence and metastasis risk and aiding clinical decision-making.

The findings were published in the international journal Advanced Science.