基于LC-MS/MS的差异血清氨基酸风险评分模型在食管鳞癌诊断中的价值分析

Value analysis of differential serum amino acid risk score model based on LC-MS/MS in the diagnosis of esophageal squamous cell carcinoma

目的基于高效液相色谱-串联质谱(LC-MS/MS)探讨食管鳞癌(ESCC)患者的差异血清氨基酸,建立基于血清差异氨基酸水平的风险评分,探究风险评分在ESCC诊断中的价值。方法采用LC-MS/MS法检测本院2018年4月至2019年3月确诊的32例ESCC患者(ESCC组)和同期年龄性别匹配的30例体检健康者(对照组)的血清氨基酸水平,比较ESCC组和对照组血清中20种非必需氨基酸和必需氨基酸水平的差异;采用受试者工作特征(ROC)曲线评估差异血清氨基酸对ESCC的诊断效能,Pearson相关性分析探讨血清氨基酸与肿瘤体积的相关性。根据差异表达氨基酸水平采用多元线性逻辑回归模型建立基于血清氨基酸含量的风险评分公式。结果与对照组相比,ESCC组的血清丙氨酸、苯丙氨酸、天冬氨酸、谷氨酸、丝氨酸和精氨酸水平升高,而脯氨酸水平降低,差异均有统计学意义(P<0.05),各差异氨基酸诊断ESCC的ROC曲线下面积(AUC)为0.647~0.738。丙氨酸、精氨酸和天冬氨酸水平与肿瘤体积呈正相关(r=0.453、0.554和0.447,P<0.05),而苯丙氨酸水平与肿瘤体积呈负相关(r=-0.347,P<0.05)。多元逻辑回归模型建立基于血清差异氨基酸水平的风险评分公式[风险评分=0.02722×苯丙氨酸+0.001433×丙氨酸-0.02425×精氨酸-0.004966×脯氨酸+0.002278×丝氨酸+0.1841×天冬氨酸-0.7159],ESCC组的风险评分高于对照组(P<0.05),风险评分诊断ESCC的AUC为0.950(敏感度和特异度分别为84.4%和100.0%),且与肿瘤体积、临床分期和淋巴结转移相关(P<0.05)。结论血清氨基酸在ESCC患者中呈现不同程度的差异表达,血清氨基酸在辅助诊断ESCC中具有一定潜在价值。

Objective To investigate the differential serum amino acids of patients with esophageal squamous cell carcinoma(ESCC)based on liquid chromatography-tandem mass spectrometry(LC-MS/MS),establish a risk score based on the levels of serum differential amino acids,and explore the value of risk score in the diagnosis of ESCC.Methods Serum amino acid levels of 32 patients with ESCC diagnosed in our hospital from April 2018 to March 2019(ESCC group)and 30 healthy people matched by age and sex at the same period(Control group)were detected by LC-MS/MS method,and differences of serum 20 non-essential amino acids and essential amino acids between ESCC group and Control group were compared.Receiver operating characteristic(ROC)curve was used to evaluate the diagnostic efficacy of differential serum amino acids for ESCC,and Pearson correlation analysis was used to explore the correlation between serum amino acids and tumor volume.According to the levels of differentially expressed amino acids,the risk scoring formula based on serum levels of amino acid was established by multiple linear logistic regression model.Results Compared with the Control group,the serum levels of alanine,phenylalanine,aspartic acid,glutamate,serine and arginine increased while the level of proline decreased in ESCC group(P<0.05).The area under ROC curve(AUC)of ESCC diagnosed by differential amino acids was 0.647-0.738.The levels of alanine,arginine and aspartate were positively correlated with tumor volume(r=0.453,0.554 and 0.447,P<0.05),while the level of phenylalanine was negatively correlated with tumor volume(r=-0.347,P<0.05).The multiple logistic regression model established a risk score formula based on serum differential amino acid levels(risk score=0.02722×phenylalanine+0.001433×alanine-0.02425×arginine-0.004966×proline+0.002278×serine+0.1841×aspartate-0.7159).The risk score of ESCC group was higher than that of the Control group(P<0.05).The AUC of risk score in the diagnosis of ESCC was 0.950 with sensitivity and specificity of 84.4%and 100.0%,which was closely related to tumor volume,clinical stage and lymph node metastasis.Conclusion Serum amino acids show differential expression pattern in patients with ESCC.Serum amino acids have certain potential in the diagnosis of ESCC.