|ECR 2018 / C-1657|
|evaluation of CT factors of transplantectomy in pancreatic graft trombosis : proposal of a radiological reading table|
a. Epidemiological data:
Among the 77 pancreatic transplanted patients in our Center, 18 presented significant thrombotic vascular complications. Thus the percentage of vascular complication on the overall transplanted population is 23.37%. Group T includes 8 patients (9.1% of the total number of pancreatic transplanted patients) and group NT includes 10 patients.
Comparison of T and NT groups highlighted a significant difference regarding cold ischaemia time, which was higher in group T (p = 0.0308). There is also a non-significant trend regarding the recipients’ BMI, which was more important in group T. At last, 6 patients in the NT group were dialysed while only 1 in the T group (p = 0.0656).
b. CT time:
The CT examinations have been conducted in 16 patients on clinical or biological suspicion of thrombosis: abdominal pain, deglobulisation, rise in blood glucose or C-peptide.
In group T, the median time between transplantation and CT examination was 3 days, the period going from 1 to 19 days (standard deviation = 6). However, 7 out of 8 patients were transplantectomised within the first 5 days after transplant.
In the NT group, the median time between transplantation and CT examination was 8 days, with a significant dispersion of measures ranging from 4 to 20 days (standard deviation = 5).
Figure 5 represents the thrombosis occurrence for each group. The risk of transplectomy is all the more important that thrombosis is diagnosed early. The very early occurrence of thrombotic complications is more likely to lead to transplantectomy (p = 0.0162).
2. Radiological analysis:
All the CT examinations were of good quality and interpretable as regards the criteria to be studied. Both readers’ radiological findings were concordant.
a) Pancreatic enhancement (table 1) (fig 4):
A significant difference in percentage of pancreatic parenchyma enhancement between the two groups (p = 0, 0239) has been highlighted. All the patients who presented a complete necrosis of the pancreatic gland had been transplantectomised (n = 4 vs 0); on the other hand we noted the same number of grafts without any enhancement anomaly (0%) in both groups (n = 2 vs 2).
b ) Number of vessels damaged involved
The number of vessels affected is significantly higher in group T with an average of 1.1 vs 3.1 in group NT (p = 0, 0005). In addition, the study of the correlation between the number of vascular segments involved and transplantectomy shows a significant correlation between the number of vessels involved and the occurrence of a transplantectomy (rho = 0, 89; p < 0.0001). The presence of at least 3 vessels involved predicts the occurrence of a transplantectomy with a 100% specificity and a 62% sensitivity. The area under the ROC curve is 0.98 (fig 5 ).
c) Thrombus location (table 2):
In group T, arterial involvement occurs in 87.5% of cases and venous involvement in 87.5% of cases. In the NT group, only a single arterial involvement is listed (10% of cases), against venous involvement in 90% of cases.
This result shows a significant association between artery involvement, and transplantectomy (p = 0, 0028), whereas there is no significant difference between the groups as regards venous involvement (p = 1).
d) Thrombosis characterization (Fig 6):
We have pointed out a significant association between complete thrombosis of a vessel (arterial or venous) and the occurrence of transplantectomy (p = 0.0008).
All the patients in group T had at least a complete thrombosis of a vessel, while in the NT group one patient only displayed a complete thrombosis. The results also show that 100% of the patients who had partial thrombosis only were able to avoid transplantectomy (Fig 7).
Thematically related posters
ECR 2018 / C-3219
Tumors invading the inferior vena cava with intracardiac extension, imaging features of a series of cases in our institution