N6-methyladenosine-related lncRNAs in combination with computational histopathology and radiomics predict the prognosis of bladder cancer

Abstract

Objectives: Identification of m6A- related lncRNAs associated with BC diagnosis and prognosis. Methods: From the TCGA database, we obtained transcriptome data and corresponding clinical information (including histopathological and CT imaging data) for 408 patients. And bioinformatics, computational histopathology, and radiomics were used to identify and analyze diagnostic and prognostic biomarkers of m6A-related lncRNAs in BC. Results: 3 significantly high-expressed m6A-related lncRNAs were significantly associated with the prognosis of BC. The BC samples were divided into two subgroups based on the expression of the 3 lncRNAs. The overall survival of patients in cluster 2 was significantly lower than that in cluster 1. The immune landscape results showed that the expression of PD-L1, T cells follicular helper, NK cells resting, and mast cells activated in cluster 2 were significantly higher, and naive B cells, plasma cells, T cells regulatory (Tregs), and mast cells resting were significantly lower. Computational histopathology results showed a significantly higher percentage of tumor-infiltrating lymphocytes (TILs) in cluster 2. The radiomics results show that the 3 eigenvalues of diagnostics image-original minimum, diagnostics image-original maximum, and original GLCM inverse variance are significantly higher in cluster 2. High expression of 2 bridge genes in the PPI network of 30 key immune genes predicts poorer disease-free survival, while immunohistochemistry showed that their expression levels were significantly higher in high-grade BC than in low-grade BC and normal tissue. Conclusion: Based on the results of immune landscape, computational histopathology, and radiomics, these 3 m6A-related lncRNAs may be diagnostic and prognostic biomarkers for BC.

Keywords

• Biomarker
• Diagnosis
• Radiomics
• Urinary bladder neoplasms
• Prognosis