全容积ADC直方图参数联合MRI征象鉴别低分化胰腺导管腺癌与G3胰腺神经内分泌肿瘤的价值

Value of whole-volume ADC histogram parameters combined with MRI features in differential diagnosis of poorly differentiated pancreatic ductal adenocarcinoma and G3 pancreatic neuroendocrine tumors

目的:探讨全容积ADC直方图参数联合MRI征象鉴别低分化胰腺导管腺癌(PDAC)和G3胰腺神经内分泌肿瘤(pNET)的价值。方法:回顾性分析经穿刺或手术病理证实的低分化PDAC 30例(低分化PDAC组)和G3 pNET 13例患者(G3 pNET组)。2组均行DWI和常规增强扫描,分析MRI征象并记录相应结果。采用Firevoxel软件计算生成ADC直方图参数,包括ADC最小值(ADC_(min))、ADC平均值(ADC_(mean))、ADC最大值(ADC_(max))、ADC第10百分位数(ADC_(10%))、ADC第50百分位数(ADC_(50%))、ADC第90百分位数(ADC_(90%))、偏度系数、峰度系数、熵。采用χ^(2)检验或Fisher确切概率法比较2组计数资料的差异,采用t检验或Mann-Whitney U检验比较2组计量资料的差异。采用ROC曲线评价差异有统计学意义的MRI征象和ADC直方图参数的诊断效能。采用单因素和多因素二元Logistic回归分析找出独立预测因子,构建鉴别诊断模型。结果:MRI示2组强化模式、上游主胰管扩张、胰周侵犯及肝脏转移情况,差异均有统计学意义(均P<0.05);2组ADC直方图参数ADC_(mean)、ADC_(min)、ADC_(max)、ADC_(10%)、ADC_(50%)、ADC_(90%)差异均有统计学意义(均P<0.05)。ROC曲线显示,MRI征象中的强化模式和ADC直方图中的ADC_(mean)的鉴别诊断效能较高。多因素二元Logistic回归构建联合模型显示,AUC为0.977,特异度100.00%,敏感度86.67%。结论:MRI征象和ADC直方图参数对低分化PDAC和G3 pNET具有一定的鉴别诊断价值,尤其当两者联合时具有较高的鉴别诊断效能。

Objective:To investigate the diagnostic value of whole-volume ADC histogram parameters combined with MRI features in differentiating diagnosis of poorly differentiated pancreatic ductal adenocarcinoma(PDAC)and G3 pancreatic neuroendocrine tumor(G3 pNET).Methods:30 patients with poorly differential PDAC and 13 patients with G3 pNET confirmed by biopsy or surgical pathology were analyzed retrospectively.All patients underwent DWI and conventional contrast-enhanced scan,MRI features were analyzed and recorded.Firevoxel software was used to calculate the whole-volume ADC histogram parameters,included minimum(ADC_(min)),mean(ADC_(mean)),maximum(ADC_(max)),10th percentile(ADC_(10%)),50th percentile(ADC_(50%)),90th percentile(ADC_(90%)),skewness,kurtosis and entropy.Chi-square or Fisher’s exact probability test were used to compare count data sets between the two groups.T-test or Mann-Whitney U test were used to compare the difference of measurement data between the two groups.ROC curves were used to analyze the diagnostic efficacy of statistically significant MRI features and ADC histogram parameters.Univariate and multivariate logistic regression analysis were conducted to find the independent variables and to build a predictive model.Results:MRI features including enhancement pattern,upstream main pancreatic duct dilation,peripancreatic invasion and liver metastasis were statistically significant between the two groups(all P<0.05).The ADC histogram parameters including ADC_(mean),ADC_(min),ADC_(max),ADC_(10%),ADC_(50%) and ADC_(90%)between the two groups were statistically significant(all P<0.05).ROC curves showed that enhancement pattern and ADC_(mean) had high value in differential diagnosis.The combined model constructed by multivariate logistic regression analysis showed the highest differential diagnostic performance.It achieved an AUC value of 0.977 with sensitivity of 100.00%and specificity of 86.67%.Conclusions:MRI features and ADC histogram parameters are valuable in differentiatal diagnosis of poorly differentiated PDAC and G3 pNET,while the combination of both has a higher diagnostic performance.