Prognostic differential subpopulation classification and immunotherapy response prediction in pancreatic cancer patients based on the gene features of necrotizing apoptosis

Kaili Liao¹Zheng Fu²Xinrui Liu³Xiajing Yu⁴Linfeng Jin⁴Jinting Cheng⁴Dongyu Yang³Kun Ai³Ziqian Liu³Daixin Guo⁴Shuai Liu⁵Xiwen Yan³Zijia Li³Mingchen Xu³Xiya Yan³Jingyi Gan⁶Zhiwen Cheng⁵Wenqing Zhu²Mingxiu Cai³Wanqian Xu³Ziying Li³Xiaozhong Wang^1*Jiasheng Xu⁷

• ¹Second Affiliated Hospital of Nanchang University, Nanchang, China
• ²The First Clinical Medical college, Nanchang University, Xuefu Road, Nanchang, Jiangxi 330006, China, Nanchang, China
• ³Queen Mary college, Jiangxi Medical College, Nanchang University, Xuefu Road, Nanchang, Jiangxi 330006, China, Nanchang, China
• ⁴School of Public Health, Nanchang University, Xuefu Road, Nanchang, Jiangxi 330006, China, Nanchang, China
• ⁵The Second Clinical Medical college, Nanchang University, Xuefu Road, Nanchang, Jiangxi 330006, China, Nanchang, China
• ⁶The forth Clinical School of Nanchang University, the South Road of Bayi Square, Nanchang, Jiangxi 330006, China, Nanchang, China
• ⁷Department of Colorectal Surgery, The Second Hospital of Zhejiang University School of Medicine, Zhejiang University, Hangzhou 310009, China, Hangzhou, China

In this study, 67 necroptosis-related genes (NRGs) were collected from the literature, and 15 genes were identified to perform the initial clustering, which classified patients into two NRG subtypes. Differentially expressed gene analysis identified 495 prognosis-related genes. These were subsequently utilized for a second clustering to define two gene subtypes. Seven key genes were then selected to construct a prognostic model, and patients were stratified into high- and low-risk groups. We found that NRGcluster A and geneCluster B largely overlapped and were predominantly classified as low-risk groups. Among the 15 NRGs, 11 showed significant expression differences between the high- and low-risk groups. We next validated the prognostic characteristics of the seven model genes in an external cohort, focusing on CHST11, the most significant prognostic marker. Spearman correlation analysis revealed that CHST11 was significantly correlated with 13 of the 15 NRGs. Furthermore, we investigated the associations between CHST11 and immune response, immune cells, and malignant epithelial cells across three single-cell datasets, and characterized its expression patterns. In addition, Mendelian randomization analysis was performed to identify genes associated with pancreatic cancer from the differentially expressed gene set, analyzing their prognostic features and correlations with CHST11 using transcriptomic data. Finally, spatial transcriptomics was applied to visualize the spatial distribution of the CHST11.