Identification of stemness subtypes and prognostic modeling in thyroid cancer: the critical role of DPYSL3 in tumor progression and immune microenvironment.

PURPOSE: The objective of this study was to develop and evaluate a novel classifier and prognostic model based on the stemness characteristics of thyroid cancer patients. METHODS: Utilizing transcriptomic data from thyroid carcinoma (THCA) patients in The Cancer Genome Atlas (TCGA) database, we calculated the stemness index (mRNAsi) using the one-class logistic regression (OCLR) method. Patients were subsequently classified into three distinct subtypes through consensus cluster analysis. RESULTS: Subtype III, characterized by its stem-like properties, exhibited significantly lower overall survival (OS) and a higher somatic mutational burden. Comprehensive analysis of the tumor immune microenvironment (TIME) in Subtype III suggested an immunosuppressive phenotype. Through the application of four machine learning algorithms and LASSO regression, we identified key genes and constructed a prognostic model based on the stemness signature. This model revealed that patients in the high-risk group had lower progression-free survival (PFS) but may benefit more from immune checkpoint blockade therapy, as indicated by TIME analysis. Functional experiments demonstrated that the stemness signature gene DPYSL3 promotes the proliferation, migration, and invasion of thyroid cancer cells and is associated with cancer stem cell properties. CONCLUSION: This study provides a new strategy for thyroid cancer immunotherapy by integrating stemness-based classification and prognostic modeling.