- Population:
This is a retrospective study of the whole population who underwent a pancreas transplantation in our University centre,
from April 2007 to December 2015. 77 pancreatic transplants were conducted during this period.
The study of the clinico-radiological records included a total of 28 patients with early vascular thrombosis within the first month following transplantation.
Distal thrombosis of the pancreatic vessel transected margin or of the mesenteric artery stump or a margined thrombus with no reduction of the vessel lumen have been considered criteria of exclusion (Fig 2).
Therefore,
we have retained 18 patients with significant thrombotic complications.
In this population of pancreatic transplanted patients with thrombotic complications,
we compared the group of patients (group T) who underwent early graft failure entailing a transplantectomy to the other group (group NT) whose vascular complication did not entail transplantectomy or graft loss to 3 months.
It should be noted that all patients received an appropriate treatment through revision surgery for thrombectomy and/or curative anticoagulation.
- Imaging technique
All imaging examinations mentioned in our study were made on requests of clinicians,
on a suspicion of post-transplant complications.
The CT examinations performed in our Center for this population were all made on a 64-detector scanner (SIEMENS SOMATOM® Definition Scanner 64 Erlangen,
Germany). The acquisition parameters were as follow: Acquisition scope: 150 cm; section: 2mm thickness; reconstruction interval: 1 mm; 120 kV,
280 mA; noise index: 15.
The acquisition protocol was triphasic covering the entire abdominal-pelvis:
-acquisition without injection.
-acquisition after injection of iodine contrast medium at arterial time with setting of an automatic detection ROI in the abdominal aorta.
-acquisition at 80 seconds after injection of iodine contrast agent.
- Epidemiological analysis:
We compared the global epidemiological data between T and NT groups.
Secondarily,
we compared the lapses of time in vascular thrombotic complication occurrences in T and NT groups respectively.
- Radiological analysis:
The imaging analysis was concomitantly conducted by two radiologists,
blinded to patient status (transplantectomised or not).
For each patient with thrombotic complication,
we studied the following radiological criteria:
1: the percentage of pancreatic necrosis : visually graded according to a semi-quantitative scale of 4 categories based on the enhancement (0%,
<50%,
>50%,
100%). These categories were chosen to simplify classification in common practice.
2: the number of thrombosed vessels: The selected vessels were the mesenteric superior artery and vein (SMA and SMV),
the splenic artery and vein (SA and SV).
3: Arterial and/or venous location of the thrombus.
4: the completeness of thrombosis : we distinguished cases of complete occlusion of the vessel lumen from those where the thrombosis was partial.
- Statistics:
For each quantitative variable a Shapiro-Wilk test was used in order to determine if our samples followed a normal distribution. Depending on this setting,
we compared the epidemiological data of both groups,
either with a Student test or with a Mann-Whitney test. We compared both groups with a Fisher test regarding the proportion of arterial and venous thrombi as well as the complete or partial status of the thrombosis. We compared the percentage differences in pancreatic graft enhancement with a χ² test. We used a Spearman correlation coefficient and built a ROC curve to analyze the relationship between the number of pancreatic vessels and transplantectomy. We compared the scanner completion times with a Mann-Whitney test. A difference was considered significant if p < 0.05. The statistical software used was MedCalc 17.4 version (MedCalc Software,
Ostend,
Belgium).