What question did this study set out to answer?

This research aims to understand how RNA modification influences inflammation and immune responses in cervical cancer. It also seeks to develop a prognostic biomarker for immunotherapy response.

March 16, 2026Open Access

Machine Learning Integration of Multiomics Data Reveals RNA Modification–Driven Inflammation–Immune Interplay and Predicts Immunotherapy Response in Cervical Cancer

Key Points

This research aims to understand how RNA modification influences inflammation and immune responses in cervical cancer. It also seeks to develop a prognostic biomarker for immunotherapy response.
Conducted integrated analysis of single-cell RNA sequencing, spatial transcriptomics, and bulk RNA sequencing data.
Developed a consensus machine learning-derived RNA methylation signature to evaluate prognostic value.
Validated findings through RT-qPCR, immunofluorescence, and structural modeling.
Significant RNA modification regulators identified in tumor epithelial cells and immune subsets within the tumor microenvironment.
The RNA methylation signature stratified patients into high- and low-risk groups, correlating with survival outcomes.
High-risk patients had enhanced immune cell infiltration but showed reduced response to PD-1/CTLA-4 blockade.

Abstract

Objective The interplay between inflammation, immunity, and RNA modification (RM) in the tumor microenvironment (TME) of cervical cancer (CC) remains poorly understood. Materials and Methods We performed integrated analysis of single‐cell RNA sequencing (scRNA‐seq), spatial transcriptomics (ST), and bulk RNA‐seq data. A comprehensive computational framework, encompassing 101 machine learning algorithms, was used to construct a consensus machine learning–derived RNA methylation signature (CMDRMS). This signature was evaluated for its prognostic value, association with immune cell infiltration, inflammatory profiles, and response to immunotherapy and pharmacotherapy. Key findings were validated through RT‐qPCR, immunofluorescence, and structural modeling via AlphaFold3. Results scRNA‐seq and ST revealed significant enrichment of RM regulators within tumor epithelial cells and specific immune subsets in the TME. The CMDRMS, comprising four hub genes (CBLLI, LARPI, NUDT3, and METTL16), effectively stratified patients into high‐ and low‐risk groups according to the mean CMDRMS score. The high‐CMDRMS group exhibited poorer overall survival, an inflamed TME with enhanced immune cell infiltration, yet a paradoxically diminished response to PD‐1/CTLA‐4 blockade, suggesting an immunosuppressive phenotype. Conversely, these patients showed potential sensitivity to PD‐L1 therapy. Drug sensitivity analysis identified 48 agents with greater efficacy in the low‐CMDRMS group. Pseudotime analysis indicated that core CMDRMS genes, particularly NUDT3, were upregulated in later differentiation stages. Experimental validation confirmed the overexpression of NUDT3 in CC tissues and cell lines, correlating with advanced clinical stages. Crucially, NUDT3 demonstrated significant coexpression and predicted structural interaction with PD‐L1. Conclusion Our study underscores the critical role of RM in shaping the inflammation–immune interplay within the CC TME and provides a novel prognostic biomarker (CMDRMS) for CC and a potential therapeutic target (NUDT3) for reversing immunotherapy resistance.

Machine Learning Integration of Multiomics Data Reveals RNA Modification–Driven Inflammation–Immune Interplay and Predicts Immunotherapy Response in Cervical Cancer

Key Points

Abstract

Cite This Study