A necroptosis gene signature predicts prostate cancer recurrence, and is linked to somatic mutation, therapeutic landscape, and immune infiltration

Necroptosis, a type of programmed necrotic cell death, has been implicated in cancer biology and therapeutics. Improved risk stratification is required for prostate carcinoma in individuals. In view of the importance of necroptosis, this work proposed a necroptosis-based genetic model for recurrence prediction, and clarified its characteristics. A least absolute shrinkage and selection operator (LASSO) regression analysis was conducted based upon transcriptome data of necroptosis genes with clinical information in the Cancer Genome Atlas (TCGA) prostate carcinoma samples, which were externally verified in the GSE116918 cohort. Somatic mutation was characterized by Maftools method. The drug sensitivity was estimated via OncoPredict algorithm. T-cell inflammation score and tumor mutational burden (TMB) score were computed for inferring immunotherapy response. CIBERSORT was adopted for scoring the infiltration of immune cell compositions. The necroptosis gene model was defined, composed of BCL2, BCL2L11, BNIP3, CASP8, CYLD, HDAC9, IDH2, IPMK, MYC, PLK1, TNF, TNFRSF1A, and TSC1. Considering external verification, this model effectively predicted recurrence-free survival, notably within one year (area under the curve (AUC) = 0.841, 0.706, 0.776, and 0.893 in the discovery, verification, total and external independent sets, respectively). Patients who had a risk score > median value were defined as high risk, while those who had risk score ≤ median value were defined as low risk. Older age, more advanced T, N, M stage, shorter disease-free survival, and more recurred/progressed statuses were found in high-risk patients (all P