Supplementary Material for: AKAP12 and IGFBP4 are prognostic factors for chronic lymphocytic leukemia

Introduction: To develop a prognostic model for chronic lymphocytic leukemia (CLL). Methods: GEO2R was used to retrieve the gene expression data of CLL and normal B cells from the Gene Expression Omnibus (GEO, GSE22529 and GSE50006 datasets) database. Perl (Practical Extraction and Report Language) was used to extract the gene expression and overall survival (OS) data of CLL patients from the CLLE-ES project in the ICGC (International Cancer Genome Consortium) database. Cox regression with Lasso was used to create and validate a prognostic model for CLL. Results: A total of 267 genes exhibited differential expression between CLL and normal B cells. Uni-Cox analysis identified 14 DEGs that correlated with overall survival (OS). Lasso multivariate evaluation demonstrated that AKAP12 and IGFBP4 are independent prognostic factors for CLL. Kaplan-Meier survival analysis revealed a significant association between the estimated risk score and survival. The area under the receiver operating characteristics (ROC) curve (AUC) was calculated to be 0.97, indicating high predictive accuracy. In addition, high AKAP12 and IGFBP4 risk score was associated with the high incidence of trisomy 12q. Conclusion: Taken together, AKAP12 and IGFBP4 are independent prognostic factors for CLL.

引言：本研究旨在构建慢性淋巴细胞白血病（chronic lymphocytic leukemia, CLL）的预后模型。方法：借助GEO2R工具从基因表达综合数据库（Gene Expression Omnibus, GEO）中提取CLL与正常B细胞的基因表达数据，所用数据集为GSE22529与GSE50006。通过Perl（实用提取与报告语言，Practical Extraction and Report Language）从国际癌症基因组联盟（International Cancer Genome Consortium, ICGC）数据库的CLLE-ES项目中，获取CLL患者的基因表达数据与总生存期（overall survival, OS）数据。采用结合Lasso正则化的Cox回归模型构建并验证CLL预后预测模型。结果：共计267个基因在CLL与正常B细胞间存在差异表达。单因素Cox分析筛选出14个与总生存期（OS）相关的差异表达基因（differentially expressed genes, DEGs）。Lasso多变量回归评估显示，AKAP12与IGFBP4为CLL的独立预后因子。Kaplan-Meier生存分析结果表明，估算所得的风险评分与患者生存状况显著相关。受试者工作特征曲线（receiver operating characteristics, ROC）下面积（AUC）为0.97，提示该模型具有较高的预测准确性。此外，高AKAP12与IGFBP4风险评分与12号染色体长臂三体（trisomy 12q）的高发生率存在关联。结论：综上，AKAP12与IGFBP4是慢性淋巴细胞白血病的独立预后因子。