Lung Cancer Stage Prediction Using Multi-Omics Data.

Area of Science:

• Oncology
• Bioinformatics
• Genomics

Background:

• Lung cancer is a leading cause of cancer mortality worldwide.
• Accurate staging is critical for effective treatment planning, differentiating between early-stage (surgery) and advanced-stage (chemotherapy, radiotherapy) disease.
• Current staging methods can be improved with advanced predictive models.

Purpose of the Study:

• To evaluate the accuracy of the random forest algorithm for lung cancer stage prediction.
• To compare prediction accuracy across different data modalities: microbiome, transcriptome, and their fusion.
• To determine if multimodal data fusion enhances lung cancer staging accuracy.

Main Methods:

• Utilized the random forest algorithm for lung cancer stage prediction.
• Analyzed data from microbiome, transcriptome, and combined (fusion) groups.
• Assessed model performance using metrics including accuracy (ACC), recall, and precision.

Main Results:

• The random forest model demonstrated varying prediction accuracies across data groups.
• Microbiome and transcriptome fusion analysis yielded the highest prediction accuracy, reaching 0.809.
• Multimodal data fusion significantly improved the accuracy of lung cancer stage prediction compared to single modalities.

Conclusions:

• Multimodal data fusion, particularly combining microbiome and transcriptome data, is a powerful approach for accurate lung cancer staging.
• The random forest algorithm, when applied to fused data, shows significant potential for clinical application in lung cancer diagnosis.
• This study highlights the importance of integrating diverse biological data for improved cancer outcome prediction.