Clustering and machine learning-based integration identify cancer associated fibroblasts genes’ signature in head and neck squamous cell carcinoma

Abstract

Background: A hallmark signature of the tumor microenvironment in head and neck squamous cell carcinoma (HNSCC) is abundantly infiltration of cancer-associated fibroblasts (CAFs), which facilitate HNSCC progression. However, some clinical trials showed targeted CAFs ended in failure, even accelerated cancer progression. Therefore, comprehensive exploration of CAFs should solve the shortcoming and facilitate the CAFs targeted therapies for HNSCC.Methods: In this study, we identified two CAFs gene expression patterns and performed the single‐sample gene set enrichment analysis (ssGSEA) to quantify the expression and construct score system. We used multi-methods to reveal the potential mechanisms of CAFs carcinogenesis progression. Finally, we integrated 10 machine learning algorithms and 107 algorithm combinations to construct most accurate and stable risk model. The machine learning algorithms contained random survival forest (RSF), elastic network (Enet), Lasso, Ridge, stepwise Cox, CoxBoost, partial least squares regression for Cox (plsRcox), supervised principal components (SuperPC), generalised boosted regression modelling (GBM), and survival support vector machine (survival-SVM).Results: There are two clusters present with distinct CAFs genes pattern. Compared to the low CafS group, the high CafS group was associated with significant immunosuppression, poor prognosis, and increased prospect of HPV negative. Patients with high CafS also underwent the abundant enrichment of carcinogenic signaling pathways such as angiogenesis, epithelial mesenchymal transition, and coagulation. The MDK and NAMPT ligand–receptor cellular crosstalk between the cancer associated fibroblasts and other cell clusters may mechanistically cause immune escape. Moreover, the random survival forest prognostic model that was developed from 107 machine learning algorithm combinations could most accurately classify HNSCC patients.Conclusion: We revealed that CAFs would cause the activation of some carcinogenesis pathways such as angiogenesis, epithelial mesenchymal transition, and coagulation and revealed unique possibilities to target glycolysis pathways to enhance CAFs targeted therapy. We developed an unprecedentedly stable and powerful risk score for assessing the prognosis. Our study contributes to the understanding of the CAFs microenvironment complexity in patients with head and neck squamous cell carcinoma and serves as a basis for future in-depth CAFs gene clinical exploration.

Creating a multifaceted prognostic model for cutaneous melanoma: the convergence of single-cell and bulk sequencing with machine learning

“…First, unifactor COX regression analysis was conducted on the TCGA dataset to identify genes strongly associated with prognosis. Subsequently, a variety of machine learning algorithms, including RSF, COXBoost, Enet, GBM, Lasso, plsRcox, Ridge, StepCox, and Survivor-SVM, were utilized individually and in combinations to construct prognostic models ( Wang Q. et al, 2023 ; Wang D. et al, 2023 ; Pei et al, 2023 ). During the construction of the model, the TCGA cohort is used as the training set and the three GEO cohort is used as the validation sets.…”