Machine Learning Tool Enables Noninvasive Diagnosis and Monitoring of Colorectal Cancer

Colonoscopy remains the gold standard for CRC diagnosis, but it is invasive, costly, and necessitates extensive bowel preparation and sedation. Recent progress in high-throughput “omics” technologies presents an opportunity for less invasive CRC diagnostic methods through the identification of biomarkers. To further enhance cancer diagnostics, researchers have now developed a machine learning-based tool capable of detecting molecular profile differences related to metabolism between patients with colorectal cancer and healthy individuals.

In a study involving biological samples from over 1,000 participants, scientists at The Ohio State University (Columbus, OH, USA) identified metabolic changes linked to disease severity and genetic mutations that elevate the risk of colorectal cancer. While additional analysis is ongoing, the resulting "biomarker discovery pipeline" holds promise as a noninvasive method for diagnosing colorectal cancer and monitoring disease progression. However, the tool is not intended to replace colonoscopy as the primary screening method for cancer, and more studies involving additional samples are needed before the pipeline can be considered for clinical application.

The research, published in the journal iMetaOmics, also demonstrates an advancement in machine learning techniques, combining two established methods to create the new platform: partial least squares-discriminant analysis (PLS-DA) for broad molecular profile differentiation, and an artificial neural network (ANN) that identifies molecules enhancing the platform's predictive capability. The researchers named the resulting biomarker pipeline PANDA, which stands for PLS-ANN-DA. For the study, the team analyzed two sets of biological data derived from blood samples: metabolites, which are products of biochemical reactions that break down food to produce energy and perform other essential functions, and transcripts, RNA readouts of DNA instructions predicting related protein changes.

Reliance on biomarkers for diagnostics across different populations can be challenging due to the wide array of conditions that affect molecular profiles in living systems. Nevertheless, this study highlights several molecular changes that show potential—though not certainty—in evaluating the presence and progression of colorectal cancer in a nationally representative group of patients. Metabolism pathways related to purines, compounds essential for DNA formation and degradation, stood out in the analysis, showing higher overall activity in cancer patients compared to healthy controls, and lower activity in those with more advanced tumor stages. The research team plans to continue analyzing metabolites associated with various biological signals to further refine the PANDA biomarker pipeline.

“We believe this is a good tool for disease diagnostics and monitoring, especially because metabolic-based biomarker analysis could also be utilized to monitor treatment effectiveness,” said Jiangjiang Zhu, co-senior author of the study and an associate professor of human sciences at The Ohio State University. “When a patient is taking drug A versus drug B, especially for cancer, time is essential. If they don’t have a good response, we want to know that as soon as possible so we can change the treatment regimen. If metabolites can help indicate a treatment’s effectiveness faster than traditional methods like pathology or protein markers, we hope they could be good indicators for doctors who are caring for patients.”