Diagnosis of Pancreatic Tumors: Comparison of MR Pancreatograp hy (MRP )and Endoscopic Retrograde Pancreatography(ERP)

Noh, Ki Suh; Seo, Jung Hoon; Kim, Myeong Jin; Chung, Jae Bok; Chung, Jae Joon; Lee, Jong Tae; Yoo, Hyung Sik

Abstract

PURPOSE: Magnetic resonance pancreatography(MRP) is a non-invasive imaging technique for visualization of the pancreatic duct system, and is similar to those obtained by means of endoscopic retrograde pancreatography (ERP). To determine the role of MRP in the diagnosis of pancreatic tumors, the diagnostic confidence and imaginal difference of MRP and ERP were compared.
MATERIALS AND METHODS
Twenty patients (13 male and 7 female, mean age 59 years) with pancreatic tumors underwent MRP and ERP. The former involved the use of a single shot fast spinecho sequence on a 1.5T system. All images were retrospectively reviewed by a radiologist and a gastroenterologist, working together. Both MRP and ERP were compared for separate visualization of the head, body and tail portion of the pancreatic duct, and scored as excellent (4), good (3), fair (2), poor (1), or no visualization (0). In addition, the overall diagnostic confidence of both modalities was graded subjectively from non-diagnoses (0) to definite information (4). The final diagnoses derived from surgical findings (n=9) or imaging findings and clinical follow-up (n=7) were as follows : pancreatic cancer (n=12), mucin-producing pancreatic cancer (n=2), mucinous ductectatic tumor (n=4), serous cystadenoma (n=2). To assess the statistical significance of difference, the paired t-test was used.
RESULTS
Mean scores of visualization of the pancreatic duct by MRP and ERP were 2.91 and 3.15 in the pancreatic head (p=NS), 3.11 and 2.18 in the pancreatic body (p=NS), and 3.07 and 1.09 in the pancreatic tail (p<0.01). The mean score of diagnostic confidence was 4.03 for MRP and 2.51 for ERP, a statistically significant difference (p <0.05). In 11 patients with obstruction of the pancreatic duct due to malignant lesions, MRP visualized the duct both proximally and distally to the site of obstruction, while ERP visualized only the distal duct to the site of obstruction. MRP was also better at defining the extent of tumor by visualization of surrounding pancreatic parenchyma. In two cases of serous cystadenoma with lack of communication between the pancreatic duct and cystic neoplasm, MRP depicted the lesion clearly whereas ERP showed no information.
CONCLUSION
MRP is better than ERP at visualizing the of pancreatic duct proximal to obstruction, assessment of tumoral extent, and diagnosis of a cystic neoplasm which does not communicate with the pancreatic duct.