Breakthrough in Pancreatic Cancer Diagnosis: Researchers Develop Novel Gene Signature

Strategic Insights -

Researchers at Stanford University have made a groundbreaking discovery in the diagnosis of pancreatic cancer, a disease that has a poor prognosis due to late-stage detection. According to a recent study published in Heliyon, the team of scientists leveraged machine learning techniques to identify a biologically relevant gene signature that can accurately differentiate pancreatic ductal adenocarcinoma (PDAC) from chronic pancreatitis and healthy controls using blood-based RNA sequencing data. The 6-gene subset identified has established biological relevance to PDAC and maintains strong classification performance, potentially facilitating earlier diagnosis and improving patient prognosis.

Key Takeaways:

  • The study identified a 15-gene signature derived from extracellular vesicle long RNA (exLR) data that successfully classified PDAC, chronic pancreatitis, and healthy controls with an F1-score of approximately 80%.
  • The 6-gene subset has established biological relevance to PDAC and maintains strong classification performance (F1-score: 71.0% in Leave-One-Out cross-validation).
  • The research proposes a promising, biologically relevant gene signature derived from blood samples for the accurate, non-invasive differentiation of PDAC and chronic pancreatitis.
  • The study leveraged machine learning techniques to analyze two distinct datasets: extracellular vesicle long RNA (exLR) and peripheral blood mononuclear cell (PBMC) RNA-Seq.
  • The researchers used the minimum Redundancy Maximum Relevance (mRMR) algorithm and support vector machine (SVM) classification to identify the 15-gene signature.
  • Francisco Carrillo-Perez, lead researcher, emphasized the critical need for informative biomarkers enabling earlier diagnosis and improved patient outcomes.
  • Additional authors for this research include Octavio Caba, Cristina Jimenez-Luna, Francisco Ortuno, Daniel Castillo-Secilla, Luis Javier Herrera, Jose Prados, Ignacio Rojas.

Statistics:

  • Approximately 80% success rate in classifying PDAC, chronic pancreatitis, and healthy controls using the 15-gene signature.
  • 71.0% success rate in classifying PDAC, chronic pancreatitis, and healthy controls using the 6-gene subset in Leave-One-Out cross-validation.
  • 2 distinct datasets used in the study: extracellular vesicle long RNA (exLR) and peripheral blood mononuclear cell (PBMC) RNA-Seq.
  • 15-gene signature identified using the minimum Redundancy Maximum Relevance (mRMR) algorithm and support vector machine (SVM) classification.

Sources:

  • Heliyon (http://www.heliyon.com)
  • NewsRx (August 1, 2025)
  • Francisco Carrillo-Perez, Stanford Center for Biomedical Informatics Research (BMIR), Stanford University, School of Medicine, Stanford, 94305-547, CA, United States.