Une nouvelle analyse sanguine pourrait révéler si le cancer du côlon reviendra avant les symptômes

A groundbreaking study suggests that analyzing the complex interactions among amino acids in a single blood sample could offer early insight into whether colon cancer might return after treatment. Researchers from the Korea Advanced Institute of Science and Technology (KAIST), Gangnam Severance Hospital, and Asan Medical Center discovered that as colorectal cancer progresses, it alters the body’s metabolism in detectable ways through the blood. This finding, published in a recent study, introduces a novel approach to predicting cancer recurrence or metastasis long before symptoms appear. The research team focused on examining how amino acids interact rather than just their individual concentrations. They analyzed the relationships among 18 circulating amino acids extracted from a small serum sample using fluorine-19 nuclear magnetic resonance spectroscopy. Their analysis revealed that as colorectal cancer advanced, the amino acid network underwent progressive reorganization. These changes reflected broader metabolic shifts happening throughout the body during tumor progression. Among the key observations was the decreasing prominence of branched-chain amino acids like valine and leucine, which are vital for muscle and energy metabolism. In contrast, glycine and serine, which are essential for DNA synthesis and rapid cell growth in cancer cells, became more abundant in the bloodstream as the disease progressed. Notably, despite being consumed by rapidly dividing cancer cells, glycine levels in the blood rose as the disease advanced. The researchers identified a glycine-centered interaction pattern within the amino acid network, interpreting this as further evidence of systemic metabolic changes associated with cancer progression. To assess clinical relevance, the team integrated these amino acid interaction patterns into machine-learning models aimed at identifying patients at higher risk of recurrence or metastasis. These models demonstrated superior accuracy compared to traditional methods. One model, incorporating both carcinoembryonic antigen (CEA), a standard blood marker for colorectal cancer, and amino acid interaction data, outperformed models relying solely on CEA or individual amino acid levels. Another model combining CEA, amino acid concentrations, and interaction-derived features showed even stronger predictive power than models based purely on amino acid measurements. Professor Ji Min Lee, who led the study, emphasized the potential of this approach to improve precision medicine. “We hope this will lead to new precision medicine technologies that can predict recurrence risk more accurately using a blood sample alone and help establish personalized treatment strategies,” he stated. Despite the promising results, some outside experts cautioned against overinterpreting the findings. Dr. Michael F. Driscoll, director of the Gastrointestinal Malignancy Program at Norton Healthcare, noted that while the approach is intriguing, further validation is necessary. He pointed out that conclusions drawn from a relatively small population-based study should not yet influence clinical practice. “This approach could be used as an add-on to other tests if it were validated in larger studies,” Driscoll explained. The challenge of translating such findings into routine clinical use remains significant. Many blood-based biomarker tests face hurdles in achieving widespread acceptance due to factors including cost, reliability, and integration into existing diagnostic protocols. Nonetheless, the study represents a step forward in leveraging metabolic profiling to enhance cancer care. As researchers continue refining these techniques, the focus remains on expanding the scope of the study and validating the approach in diverse patient populations. If successful, this method could eventually enable earlier intervention and more tailored treatment plans for individuals at risk of colorectal cancer recurrence.