Subtype-specific expression and prognostic significance of creatine kinase family genes in non-small cell lung cancer: a TCGA-based integrated analysis

Background Non-small cell lung cancer (NSCLC), primarily comprising lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC), remains a leading cause of cancer-related mortality worldwide. While creatine kinase (CK) enzymes are central to cellular energy homeostasis, their subtype-specific transcriptional dysregulation and prognostic relevance in NSCLC remain poorly characterized. Methods RNA-seq STAR counts and clinical metadata for TCGA-LUAD and TCGA-LUSC were retrieved using TCGAbiolinks. Matched tumor (TP) and normal tissue (NT) samples were analyzed for paired differential expression analysis (DEA). Data processing included edgeR filtering, TMM normalization, and limma-voom modeling for LUAD, LUSC and pooled cohorts. The analysis focused on five CK family genes: CKM , CKB , CKMT1A , CKMT1B , and CKMT2 . Clinical characteristics and ESTIMATE-based tumor purity were incorporated into univariate and multivariate Cox regression models. Kaplan-Meier survival analysis was performed to evaluate the prognostic impact of key CK members. Results The paired cohort consisted of 109 patients (58 LUAD and 51 LUSC). DEA and volcano plots revealed distinct subtype-dependent expression patterns: While CKMT1A and CKMT1B were consistently upregulated across all cohorts, their fold-change was markedly higher in LUSC than in LUAD. Conversely, subtype-specific dysregulation was observed for other members: CKM was significantly upregulated specifically in LUAD, whereas CKMT2 was significantly downregulated only in LUSC, and CKB exhibited significant downregulation in LUAD. Heatmap analysis further supported heterogeneous expression behavior across the five genes. Multivariate Cox regression identified CKM as the only CK member with independent prognostic significance for overall survival in LUAD and the pooled cohort. Kaplan-Meier analysis demonstrated significant survival separation between CKM high- and low-expression groups in LUAD (Log-rank P  = 0.024) and the pooled cohort ( P  = 0.0026), whereas no significant difference was observed in LUSC. Conclusions CK family genes exhibit subtype-specific transcriptional alterations in NSCLC, and CKM expression serves as a robust prognostic marker, particularly in LUAD. These findings suggest that creatine kinase-mediated metabolic pathways may contribute to subtype-specific outcome heterogeneity and highlight CK genes as potential targets for refined clinical risk stratification.