Abstract
Ovarian cancer (OC) is a disease with difficult early diagnosis and treatment and poor prognosis. OC data profiles were downloaded from The Cancer Genome Atlas. Eight key fatty acid metabolism-related long non-coding RNAs (lncRNAs) were finally screened for building a risk scoring model by univariate/ multifactor and least absolute shrinkage and selection operator (LASSO) Cox regression. To make this risk scoring model more applicable to clinical work, we established a nomogram containing the clinical characteristics of OC patients after confirming that the model has good reliability and validity and the ability to distinguish patient prognosis. To further explore how these key lncRNAs are involved in OC progression, we explored their relationship with LUAD immune signatures and tumor drug resistance. The structure shows that the risk scoring model established based on these 8 fatty acid metabolism-related lncRNAs has good reliability and validity and can better predict the prognosis of patients with different risks of OC, and LINC00861 in these key RNAs may be a hub gene that affects the progression of OC and closely related to the sensitivity of current OC chemotherapy drugs. In addition, combined with immune signature analysis, we found that patients in the high-risk group are in a state of immunosuppression, and Tfh cells may play an important role in it. We innovatively established a prognostic prediction model with excellent reliability and validity from the perspective of OC fatty acid metabolism reprogramming and lncRNA regulation and found new molecular/cellular targets for future OC treatment.
Introduction
Ovarian cancer (OC) remains a "silent killer" in oncology, often diagnosed only in advanced stages when the five-year survival rate drops below 30%. A critical aspect of OC progression is metabolic reprogramming, where cancer cells alter pathways like fatty acid metabolism (FAM) to support rapid growth and survival. While the importance of FAM is recognized, the regulatory influence of long non-coding RNAs (lncRNAs) on these metabolic processes has remained largely unexplored. This study aims to bridge this gap by identifying FAM-related lncRNAs that can serve as robust prognostic biomarkers and clinical decision-making tools.
Methodology
The study utilized a comprehensive bioinformatics pipeline to analyze data from 364 OC patients sourced from The Cancer Genome Atlas (TCGA-OV). The researchers first identified 92 FAM-related genes using the GeneCards database and performed correlation analyses to find associated lncRNAs. To ensure clinical relevance, drug sensitivity was predicted using the Genomics of Drug Sensitivity in Cancer (GDSC) database, and immune subtypes were validated via TIMER2.0.
To ensure the prognostic model's accuracy and avoid overfitting, the team employed a combination of univariate Cox regression and Least Absolute Shrinkage and Selection Operator (LASSO) Cox regression. This process narrowed down hundreds of potential candidates to a final signature of eight key lncRNAs: AC080013.1, GS1-124K5.4, AC099518.1, AC011595.1, LINC00861, AC027279.1, AC145343.1, and AL021707.1.
Results and Key Insights
The resulting 8-lncRNA risk scoring model demonstrated high predictive power, achieving an Area Under the Curve (AUC) of 0.770 for 1-year survival. Patients stratified into the low-risk group showed significantly better overall survival outcomes compared to the high-risk group (P < 0.001).
Beyond survival prediction, the study revealed significant insights into the tumor microenvironment. High-risk patients were found to be in a state of immunosuppression, characterized by lower infiltration of Follicular B helper T cells (Tfh). Furthermore, LINC00861 emerged as a central "hub gene." Its downregulation in advanced stages suggests it may play a protective role, potentially acting through the PTEN/AKT/mTOR pathway. Drug sensitivity analysis also indicated that high-risk patients might be more responsive to certain non-standard chemotherapeutic agents, providing potential alternatives for refractory cases.
Discussion and Conclusion
By integrating the 8-lncRNA risk score with clinical factors like age and tumor grade, the researchers developed a clinical nomogram with excellent calibration. This tool offers a more personalized approach to predicting OC patient outcomes. While this study is primarily computational, it identifies specific molecular targets—particularly LINC00861 and Tfh cells—for future experimental validation. These findings underscore the potential of targeting metabolic-immune cross-talk as a novel therapeutic strategy in ovarian cancer.
Summary updated