GCdiscrimination

GCdiscrimination classifies gastric cancer (GC) by analyzing 58 routine blood biochemical indices from liquid biopsies to discriminate GC from other stomach diseases, other cancers, and healthy individuals.

Key Features:

• Integrated markers: Utilizes 58 routine blood biochemical indices as candidate markers for GC identification.
• Discrimination model: Constructs a random forest model that ranks and selects the top 17 indices, comprising eight routine blood indices (MO%, IG#, IG%, EO%, P-LCR, RDW-SD, HCT, RDW-CV) and nine biochemical indices (TP, AMY, GLO, CK, CHO, CK-MB, TG, ALB, γ-GGT).
• Performance metrics: Reports cross-validation performance with sensitivity 0.9067, specificity 0.9216, total accuracy 0.9138, and area under the curve (AUC) 0.9720.
• Rapid assessment: Enables rapid, real-time preliminary assessment of gastric cancer prior to invasive diagnostic procedures such as histological analysis or gastroscopy.
• Clinical correlations: Identifies correlations between GC and routine blood biochemical parameters to inform potential mechanistic insights, prevention programs, and surveillance management strategies.

Scientific Applications:

• Screening and preliminary diagnosis: Non-invasive classification of GC for preliminary screening using blood biochemical indices.
• Differential diagnosis: Discriminates GC from other stomach diseases, different cancers, and healthy individuals.
• Biomarker discovery: Prioritizes candidate blood-based biomarkers (including MO%, IG#, IG%, EO%, P-LCR, RDW-SD, HCT, RDW-CV, TP, AMY, GLO, CK, CHO, CK-MB, TG, ALB, γ-GGT) for further investigation.
• Surveillance and prevention: Supports development of surveillance management strategies and prevention programs informed by blood biochemical parameter correlations with GC.

Methodology:

Model training and feature ranking were performed using a random forest method on 58 routine blood biochemical indices, with model evaluation by cross-validation reporting the stated sensitivity, specificity, accuracy, and AUC.