目的 探讨颗粒酶B（GZMB）和C-X-C基序趋化因子配体9（CXCL9）在胃癌组织中的表达特征、与肿瘤免疫微环境（TME）的关联及其临床预后意义。 方法 利用TIMER、GEPIA、UALCAN等公共数据库进行生物信息学分析，验证GZMB与CXCL9在泛癌及胃癌中的表达情况；收集2018—2019年南阳市第一人民医院89例行根治术的胃癌患者的癌及癌旁组织标本，采用实时荧光定量PCR（qRT-PCR）检测GZMB与CXCL9的mRNA表达水平，分析其表达与患者临床病理特征的关系，通过TIMER和GSCA数据库分析其与免疫细胞浸润的相关性；采用受试者工作特征曲线（ROC）评估其诊断价值，通过Kaplan-Meier Plotter数据库以及后续的随访进行生存分析。 结果 数据库分析及qRT-PCR结果均显示，GZMB与CXCL9在胃癌组织中的表达显著高于癌旁正常组织（P < 0.05）。临床关联分析表明，GZMB、CXCL9 mRNA表达的高低与性别、年龄、病理分型、肿瘤分化程度、TNM分期、肿瘤最大径以及淋巴结转移均无关（P > 0.05）。免疫浸润分析显示，两者均与CD8? T细胞、树突状细胞浸润呈显著正相关（P < 0.05），与B细胞呈负相关（P < 0.05）。ROC曲线分析显示，GZMB与CXCL9联合检测诊断胃癌的曲线下面积（AUC）为（0.890，95%CI = 0.843 ~ 0.936），明显高于GZMB与CXCL9单项检测的AUC（0.832，95%CI = 0.772 ~ 0.891；0.782，95%CI = 0.715 ~ 0.850）。生存分析显示，GZMB与CXCL9高表达组患者的总生存期显著优于低表达组（P < 0.05）。 结论 GZMB与CXCL9在胃癌中高表达与抗肿瘤免疫细胞浸润密切相关，是潜在的胃癌诊断生物标志物，且两者高表达提示患者预后更好。

Objective To investigate the expression characteristics of granzyme B （GZMB） and C-X-C motif chemokine ligand 9 （CXCL9） in gastric cancer （GC） tissues， explore their association with the tumor immune microenvironment （TME）， and evaluate their clinical prognostic significance. Methods Bioinformatic analyses were conducted using public databases （TIMER， GEPIA， UALCAN） to validate the expression levels of GZMB and CXCL9 across various cancer types， with a focus on GC. Paired tumor and adjacent normal tissue samples were collected from 89 GC patients who underwent radical gastrectomy at Nanyang First People's Hospital between 2018 and 2019. The mRNA expression levels of GZMB and CXCL9 were measured using quantitative real-time PCR （qRT-PCR）， and their associations with clinicopathological characteristics were statistically analyzed. Immune cell infiltration levels were estimated using the TIMER and GSCA databases. Diagnostic performance was evaluated through receiver operating characteristic （ROC） curve analysis， while survival outcomes were assessed using data from the Kaplan-Meier Plotter database and patient follow-up records. Results Both database analysis and qRT-PCR results demonstrated that GZMB and CXCL9 expression levels were significantly elevated in GC tissues compared to adjacent non-tumor tissues （P < 0.05）. Clinical correlation analysis revealed no significant associations between mRNA expression levels of GZMB and CXCL9 and clinicopathological parameters， including gender， age， pathological type， tumor differentiation， TNM stage， tumor size， or lymph node metastasis （P > 0.05）. Immune infiltration analysis indicated that both genes were significantly positively correlated with CD8⁺ T cell and dendritic cell infiltration （P < 0.05）， while showing a negative correlation with B cell infiltration （P < 0.05）. ROC curve analysis showed that the combined detection of GZMB and CXCL9 yielded an AUC of 0.890 （95%CI： 0.843 ~ 0.936）， which was higher than that of GZMB alone （AUC = 0.832， 95%CI： 0.772 ~ 0.891） or CXCL9 alone （AUC = 0.782， 95%CI： 0.715 ~ 0.850）. Survival analysis further revealed that patients with high expression of GZMB and CXCL9 had significantly improved overall survival compared to those with low expression （P < 0.05）. Conclusion GZMB and CXCL9 are highly expressed in GC and are strongly associated with the infiltration of anti-tumor immune cells. These molecules represent promising diagnostic biomarkers for GC， and their elevated expression is correlated with a more favorable prognosis in patients.