Novel serous effusion-related risk models and biomarkers for predicting prognosis in T-cell lymphoma patients.

T-cell lymphomas (TCLs) are a cluster of lymphoproliferative diseases with high heterogeneity, which lack accurate prognostic models and standard treatment regimen at present. Serous effusion (SE) is a relatively common manifestation and poses more challenges for risk stratification in TCLs. In this study, entire of 518 newly diagnosed TCLs patients were included. SE was found to be tightly correlated to clinical characteristics and prognosis in TCL patients, and SE volume (SEV) > 1000 ml was identified as a potential prognostic factor. Novel AEBS risk model, including age > 60, ECOG PS > 1, β2-microglobulin (BMG) > 3.0 mg/L and SEV > 1000 ml, which exerted superior efficacy for risk stratification compared to the current risk systems in TCL patients with SE. Besides, multiple RNA-seq datasets were used for the identification and function analysis of SE-related genes (SERGs). TCL patients in different SERGs-associated subgroups exhibited discrepancy in the infiltration of immunocytes and the expression of immune checkpoints. SERGs signature, including HIF1A, FERMT2, NFATC1 and COL1A1, was established and demonstrated to have distinguishing capacity for predicting prognosis in TCL patients. Moreover, immunohistochemistry revealed that SE-related molecule HIF1A was reductively expressed and related to inferior prognosis in TCL patients, especially in SE group. Pan-cancer analysis found HIF1A expression was decreased in several tumors, and chemosensitivity analysis revealed that HIF1A was associated with sensitivity of several anti-tumor drugs, such as Sorafenib, Navitoclax, and Venetoclax. Our findings provide evidence for identifying high-risk population and facilitating individualized treatment in TCL patients with SE.