Multi-omics clustering analysis carries out the molecular specific subtypes of thyroid carcinoma: implicating for the precise treatment strategies
Zhenglin Wang,Qijun Han,Xianyu Hu,Xu Wang,Rui Sun,Siwei Huang,Wei Chen,Wang,Z.,Han,Q.,Hu,X.,Wang,X.,Sun,R.,Huang,S.,Chen,W.
DOI: https://doi.org/10.1101/2024.02.25.24303184
2024-02-28
MedRxiv
Abstract:Background Thyroid cancer is the most prevalent endocrine malignancy, Recent classifications highlight the importance of molecular characteristics in TC, including BRAF, TERT, and RET fusion gene mutations, which are crucial for diagnosis, prognosis, and targeted therapy. This study aims to explore molecular subtypes of TC to identify new biomarkers and improve patient selection for targeted therapies. Methods This study utilized multi-omics data from the TCGA-THCA dataset and additional cohorts (GSE29265, GSE33630, GSE54958, GSE65074) involving a total of 539 patients. Various data types, including DNA methylation, gene mutations, mRNA, LncRNA, and miRNA expression, were analyzed. The study employed consensus clustering algorithms to identify molecular subtypes and used various bioinformatics tools to analyze genetic alterations, signaling pathways, immune infiltration, and responses to chemotherapy and immunotherapy. The statistical significance was established at P < 0.05. Results Two prognostically relevant thyroid cancer subtypes, termed CS1 and CS2, were identified. CS2 was associated with a poorer prognosis of shorter progression-free survival times (P < 0.001). CS1 exhibited higher copy number alterations but lower tumor mutation burden (TMB) than CS2. Notably, CS2 showed higher TMB and cytolytic activity scores, suggesting a potential for higher immunogenicity. Different pathway activations were observed between subtypes, with CS2 showing activation in cell proliferation and immune-related pathways. Drug sensitivity analysis indicated CS2's higher sensitivity to cisplatin, doxorubicin, paclitaxel, and sunitinib, whereas CS1 was more sensitive to bicalutamide and FH535. The different activated pathways and sensitive to drugs for subtypes were further validated in external cohort. After dimensionality reduction, five genes of CXCL17, LCN2, MUC1, SERPINA1, and SLC34A2 were validated that can distinguish subtypes across pan-cohorts. 24 paired tumor and adjacent normal tissues by immunohistochemical staining further show the prognostic value of CXCL17 for advanced thyroid cancer. Conclusion The study revealed two distinct molecular subtypes of thyroid cancer with significant implications for prognosis, genetic alterations, pathway activation, and treatment response. These findings underscore the potential of multi-omics approaches in enhancing personalized medicine in thyroid cancer.