Machine Learning and Pan‐Cancer Analysis of Tertiary Lymphoid Structures: A Potential Target for Survival and Drug Treatment

Abstract

Tertiary lymphoid structure (TLS) is considered to be closely related to tumor prognosis and immune response. Therefore, it has predictable clinical significance for survival prognosis and cancer drug treatment to build TLS signature of pan‐cancer. The transcriptome sequencing data of The Cancer Genome Atlas pan‐cancer analysis are obtained from UCSC Xena, including 33 cancer types (N = 10 066). A 15 TLS genes‐based TLSscore is constructed using the Random Forest and Cox regression analyses. Then, according to TLS signature, patients are classified into low‐ and high‐risk groups. Then, the differences between the two groups in terms of survival prognosis, functional enrichment, immune infiltration, and drug sensitivity are further explored. On the basis of machine learning algorithm, TLS signature has satisfactory prediction performance for the prognosis of pan‐cancer patients. A nomogram with high predictive performance is formulated by incorporating TLS signature and clinical features. The function enrichment analysis suggests that TLS signature may be related to the key pathways of immunotherapy for pan‐cancer, and shows significant differences among groups at the immune checkpoint genes. In general, a novel TLSscore‐based model is established through comprehensive analysis of TLS genes, which can accurately predict the clinical prognosis and drug sensitivity of pan‐cancer.