慢性胰腺炎模型胰泌素刺激动态MR胰胆管成像表现及与病理程度分级的相关性

Secretin-stimulated MR cholangio-pancreatography and pathological correlative study in a swine obstructive chronic pancreatitis model

目的探讨慢性胰腺炎（CP）胰泌素刺激MR动态胰胆管成像（sMRCP）表现及其与病理严重程度分级的相关性。方法将39头小型猪采用随机数字表法分为实验组（27头）和对照组（12头），对实验组猪采用不全结扎主胰管法建立CP模型，对照组猪不结扎胰管。术后4、8、12周分别随机（数字表法）选取9头实验组、4头对照组猪行常规MRI平扫及sMRCP。静脉注射胰泌素，剂量为0．6μg／kg，分别于注射前和注射后1、3、5、7、9、11min行sMRCP扫描。MRI检查结束后处死动物获取胰腺组织行病理检查，并按照病理严重程度将实验猪分为正常组、轻度CP组和中重度CP组。观察MRI表现，并在动态sMRCP图像上观测胰管基础直径、胰管最大径、胰管最终直径（注射胰泌素后11min的直径）和胰管扩张达峰时间，计算胰管最大扩张率和胰管最终变化率（注射胰泌素后11min的变化率），并评估十二指肠充盈程度分级。不同病理程度组间动态sMRCP图像上述指标的比较采用方差分析、x^2检验和U检验，相关性分析采用Spearman等级相关检验，并采用ROC评价不同参数诊断CP的效能。结果CP建模成功并获得满意sMRCP图像的实验猪纳入研究，正常组11头，轻度CP组8头，中重度CP组11头。正常组、轻度CP组和中重度CP组间主胰管基础直径分别为（1．56±0．46）、（2．95±1．17）、（7．41±1．91）mm，胰管最终直径为（1．49±0．31）、（2．96±1．17）、（7．37±1．90）mm，胰管最大径为（2．39±0．43）、（3．91±1．27）、（7．86±1．87）mm，胰管最大扩张率中位数为42．10％、34．85％和6．58％，十二指肠充盈程度分级评分中位数为3、3、2分，不同病理严重程度分级组间差异均有统计学意义（P均〈0．01），主胰管基础直径、胰管最终直径、胰管最大径、胰管最大扩张率和十二指肠充盈程度分级评分和病理分级间均具有相关性（r值分别为0．89、0．92、0．90、-0．85、-0．66，P均〈0．01）。各组问胰管扩张达峰时间和胰管最终变化率间的差异无统计学意义（P均〉0．05），和病理严重程度间不具有相关性（P均〉0．05）。主胰管基础直径、胰管最终直径、胰管最大径、胰管最大扩张率、十二指肠充盈程度分级评分鉴别正常组和轻度CP组的ROC下面积分别为0．915、0．977、0．926、0．778和0．472，鉴别轻度和中重度CP组的ROC下面积分别为0．966、0．966、0．960、1．000和0．915。结论CP的sMRCP表现具有特征性，可在体无创评估CP病理程度分级。

Objective To investigate the correlativity between secretin-stimulated magnetic resonance cholangiopancreatography （sMRCP） findings and pathological severity in a swine chronic pancreatitis （CP） model. Methods Thirty-nine swine were divided randomly into control group （n= 12） and experimental group （n= 27）. In experimental group, the main pancreatic duct （MPD） was incompletely ligated to establish the model of obstructive CP. In control group, laparotomy was performed but without ligating the MPD. At the 4th, 8th and 12th week after modeling, one third swine of each group were undergone a series of dynamic sMRCP scans before （0 rain） and at 1, 3, 5, 7, 9, 11 rain after administration of secretin （0.6 μg/kg）. And the MPD diameter and duodenum filling （DF） degree were measured. All survivals were sacrificed to pathological examination including HE and Van Gieson staining for histopathological grading. According to pathological severity, swine were divided into normal group, mild CP group and moderate to severe CP group. MRI features and indexes, including baselined diameter （BD）, end diameter （ED）, maximum diameter （MD）, the largest expansion rate （LER）, time to peak （Tpeak） and end change rate of pancreatic duct and duodenal filling （DF） scores were measured. The relationships between pathological grading and sMPCP indexes were analysed. The comparison of sMRCP data among the 3 groups were used ariance analysis, x^2 test and U test. Correlations between sMPCP indexes and pathological severity were tested using Spearman rank correlation coefficients. The diagnostic efficiency of sMRCP indexes were evaluated by ROC method. Results （1） In experimental group, 22 CP models were established and 19 CP swine （mild CP, n= 8; moderate and severe CP, n=11） were performed sMRCP successfully. Eleven swine in normal group were obtained satisfactory MRCP images. （2） sMRCP results： BD of 3 groups were （1.56±0.46）,（2.95±1.17）,（7.41±1.91） mm, respectively. ED were （1.49±0.31）,（2.96±1.17） and （7.37±1.90） mm, respectively. MD were （2.39±0.43）,（3.91±1.27） and （7.86±1.87）mm, respectively. The median of LER were 42.10%, 34.85% and 6.58%, respectively. The median of DF scores were 3, 3, 2, respectively. The differences of above indexes have statistically significance （P values were all〈0.01）. There were correlation between sMRCP indexes （BD, ED, MD, LER and DFscores） and pathological severity （ r values were 0.89, 0.92, 0.90, - 0.85, - 0.66, respectively and P values were all〈 0.01 ）. Tpeak and end change rate of pancreatic duct had no significant differences （P values were〉0.05）,and no correlation with pathological severity（P values were all〉0.05）.For differential diagnosis between normal and mild CP, the area under ROC of BD, ED, MD, LER and DFscores were 0.915, 0.977, 0.926, 0.778 and 0.472, respectively and differential diagnosis between mild CP and moderate to severe CP group, the area under ROC were 0.966,0.966,0.960,1.000 and 0.915, respectively. Conclusions sMRCP findings of CP have characteristics and could be used for in vivo evaluation on the CP pathologic grades.