Machine Learning-Based Biomarker Discovery from Serum Trace Elements and Biochemical Parameters in Patients with...

Area of Science:

• Biochemistry
• Immunology
• Medical Diagnostics

Background:

• Nasal polyps (NP) are benign growths linked to chronic inflammation, impacting patient quality of life.
• Understanding the biochemical changes in NP is crucial for identifying diagnostic markers.

Purpose of the Study:

• To evaluate inflammation, oxidative stress, and trace element homeostasis in NP patients.
• To identify potential non-invasive diagnostic biomarkers for nasal polyps.

Main Methods:

• Serum trace elements (zinc, copper, selenium) and biochemical parameters (WBC, glucose) were measured.
• Advanced statistical analyses including OPLS-DA and machine learning (logistic regression, SHAP) were employed.
• Receiver operating characteristic (ROC) analysis was used to assess biomarker discriminative power.

Main Results:

• NP patients showed higher white blood cell count (WBC), copper (Cu), and glucose levels compared to controls.
• Significantly lower serum zinc (Zn) levels and Zn/Cu ratio were observed in NP patients.
• Optimized logistic regression achieved 100% classification accuracy, with Zn, Cu, and glucose as key features.

Conclusions:

• Inflammation, trace element imbalance (Zn, Cu), and metabolic alterations are detectable in NP patients.
• Serum Zn, Cu, Zn/Cu ratio, glucose, and inflammatory indices show promise as non-invasive diagnostic biomarkers for NP.
• Further validation in larger cohorts is needed for clinical application.