Physicians and scientists, interested in MRE of the abdomen

An accurate diagnosis of pancreatic fibrosis is clinically important and may have potential for staging chronic pancreatitis (CP)¹. Endoscopic ultrasound elastography (EUS-EG) has recently shown potential to diagnose pancreatic fibrosis. However, its invasiveness, sampling limitation and a small, but not insignificant, morbidity cannot be ignored. MR elastography (MRE) can provide reproducible stiffness measurements at 40 Hz throughout the pancreas². Thus, this study was aimed at assessing the correlations of pancreatic stiffness with different pathological findings in CP (fibrosis, inflammation and fat deposition), and the potential value of MRE in staging pancreatic fibrosis.
This study was approved by our Institutional Review Board and written informed consent was obtained from all patients. 56 consecutive patients (32 men, 24 women) with pancreatic or bile duct tumors were recruited in this retrospective study between Jan. 2014 and Oct. 2015. The mean age was (47.6± 17.4) years [range: (23-78) years]. The patients all underwent pancreatectomy, and histopathologic diagnosis was obtained for the nontumoral parenchyma distal to the tumor. Twenty patients with inadequate specimens for assessment of nontumoral pancreatic parenchyma were excluded (e.g. tumor at tail or severe parenchymal atrophy). MRE using a multislice echo planar imaging (EPI) sequence with a three-dimensional (3D) wave field inversion algorithm² has been used for providing complete pancreatic stiffness maps on a 3.0T GE scanner (Signa HDX 3.0T system; GE Healthcare, Milwaukee, WI). The imaging parameters were as follows: frequency=40Hz; TR/TE = 1375/38.8ms; phase offsets = 3; FOV = 40 cm; acquisition matrix = 96×96; number of signal averages = 1; frequency-encoding direction = RL; parallel imaging acceleration factor = 3; number of slices = 32; slice thickness = 3.5 mm. The stiffness of the nontumoral parenchyma was calculated by averaging the stiffness from 2-3 ROIs (one ROI for each slice) in the parenchyma distal to the masses, avoiding boundaries and large vessels. Histological fibrosis was graded into 4 categories (normal: F0, mild fibrosis: F1, marked fibrosis,F2; and severe fibrosis:F3)^3,4. Grading criteria for fat deposition were as follows: L0: 0%-10% deposition, L1:10%-20%, L2: 20%-30%, and L3: greater than 31%. The grade of inflammation was described as absent (A0), focal (A1) or generalized infiltration (A2) by inflammatory lymphocytes.A receiver operating characteristic (ROC) curve analyses was used for prediction of pancreatic fibrosis stages.The fibrosis stage of parenchma included 14 cases in F0, 11 cases in F1, 4 cases in F2 and 7 cases in F3(Fig.1). The parenchymal stiffness showed a positive correlation with fibrosis and inflammation grades(r = 0.832, 0.616, respectively; both P< 0.001), and a trend of negative correlation with fat deposition (r = -0.301, P= 0.075). The mean stiffness showed statistical differences in the following groups of fibrosis stages: F3 vs. F0, F1 and F2 group (all P< 0.001), and F2 vs. F0, F1 group (P< 0.001 and P= 0.007), but not between F0 and F1 group (P= 0.267). Both fibrosis (P< 0.001) and inflammation (P= 0.014) contribute to the higher stiffness of pancreatic parenchyma (Stepwise multiple regression tests: r²=0.728), excluding fat deposition (P= .082). The cutoffs, sensitivity and specificity for different fibrosis stages were shown in table 1, Fig.2.The average stiffness of the fat tissue surrounding pancreas was 0.72±0.12 kPa.Both fibrosis and inflammation cause an increase in the stiffness measurement in liver. In this study, we proved similar results in CP.Both pancreatic fibrosis and inflammatory infiltrations contribute to stiffer pancreas, and pancreatic fibrosis was the predominant factor. Although the fat tissue has lower stiffess than normal pancreas, fat deposition did not significantly correlate with softer pancreatic tissue in this study. Pancreatic texture was found firm or hard in cases of usual CP and obstructive CP due to pancreatic or biliary tumors. However, histological findings of usual CP are difficult to obtain because there are few patients underwent surgery¹. Therefore, by examining the obstructive pancreatitis distal to tumor, we can also evaluate the usefulness of this diagnostic medality for various degrees of fibrosis. Although there was significant overlap in the stiffness between F0 and F1 stage, the stiffness increased sharply in F1-F2-F3 stage, with AUCs of 0.97(≥F2) and 1.0(=F3). Beside fibrosis, lipomatosis and inflammation discussed in this study, the cell density and edema might also be considered histological factors that affect the stiffness of the pancreatic parenchyma3, calling for further studies in a larger cohort of patients.Pancreatic MR elastography is promising in grading pancreatic fibrosis, which might be useful for objective and early diagnosis of CP.