Dynamic Magnetic Resonance Imaging (Dmri) of the Pancreas As a Predictor of Anastomotic Leakage After Pancreatic Resections
Marco Niedergethmann*1, Dietmar J. Dinter2, Stefan Post1
1Dept. of Surgery, University Hospital Mannheim, Mannheim, Germany; 2Dept. of Radiology, University Hospital Mannheim, Mannheim, Germany
Background: The degree of pancreatic fibrosis is a main factor for leckage after pancreatico-jejunostomy. “Soft” and “firm” pancreatic tissue types are characterized by different perfusion behaviour in DMRI. In order to identify risk factors for postoperative pancreatic anastomotic leakage and in order to assess the role of pancreatic DMRI in detection of this complication and in prediction of pancreatic texture a cohort of patients was retrospectively reviewed.
Methods: Between 2000 and 2006 a total of 107 consecutive patients were examined by means of a standardized DMRI protocol (1.5 tesla MRI): 1) morphological T1 and T2 sequences, 2) transversal T1 with fat suppression (5 mm layer thickness) at 25 and 60 seconds after i.v. application of Gd-DTPA (Magnevist®, 0,1 mmol/kg body weight), 3) T1 sequences after Gd-DTPA application. The time intensity curve on DMRI was measured in the aorta, the pancreas, and the muscle tissue. For all patients with a standardized contrast medium curve in the aorta (n=72) a muscle-normalized time intensity curve was calculated. The time intensity curves were classified in three groups: rapid increase (>1.1, early-arterial value > portal-venous value), intermediate (0.9-1.1), and slow increase (<0.9). All patients received a pancreatico-jejunostomy (duct-to-mucosa). The DMRI data was correlated with prospectively acquired clinical data.
Results: Leakage of the pancreatico-jejunostomy occurred more frequently (27%, p<0.05) in patients with a time intensity curve >1.1 (“soft” pancreas, n=37) compared to those with intermediate (0.9-1.1, n=11) and slow curves (<0.9, n=24, “firm” pancreas). Furthermore, patients with a rapid increase (>1.1) had significantly more abdominal complications such as abscess, delayed gastric emptying, as well as a pre-existing diabetes (p<0.05).
Discussion: DMRI with time intensity curve calculation provides reliable information for prediction of the pancreatic texture. A rapid increase of the time intensity curve correlates with anastomotic leakage and further abdominal complications after pancreatic head resection.