Tumor Lactate Metabolism and Immune Suppression Multi-Omics Dataset

A multi-omics dataset integrates gene signatures, digital pathology, and clinical outcomes to study lactate metabolism in cancer. It includes a curated 59-gene lactate signature and deep learning predictions from H&E whole-slide images. The data links lactate activity to tumor proliferation, immune suppression, and therapeutic resistance across TCGA, GEO, and a real-world HNSCC cohort.

Use Cases

• Validate the 59-gene lactate signature against clinical outcomes like immunotherapy or radiotherapy response.
• Correlate deep learning model predictions from H&E images with protein expression levels of LDHA and MCT1.
• Analyze associations between lactate activity states (LAC_H vs. LAC_L) and measures of immune infiltration.
• Assess model generalizability by comparing performance metrics (AUC 0.73–0.89) across 12 TCGA cancer types.

Strengths

• Integrates multi-modal data including a defined 59-gene signature, digital pathology, and clinical validation.
• Externally validated in an independent real-world SAZHU-HNSCC cohort with immunohistochemistry confirmation.
• Deep learning model achieved AUC scores between 0.73 and 0.89 across multiple cancer types.

Limitations

• Specific row and column counts are unknown, limiting assessment of dataset scale and feature completeness.
• As a 1.8 MB XLSX file, the dataset is small and likely contains aggregated summary results, not raw underlying data.
• Focus on HNSCC and 12 TCGA cancers may limit direct applicability to other, less-studied cancer types.

Provenance

Source

figshare, authored by Bohai Feng.

Collection Method

Curated from TCGA, GEO, and single-cell RNA-seq datasets, integrated with spatial and computational analyses.

Time Range

null

Freshness

Last updated March 2026.

Geography

null