Application Of¹⁸F-FDG PEG-CT In Pancreatic Cancer

Objective To study18F-FDG uptake in pancreas and influencing factors in healthy population for cancer screening.Methods From January2007to December2011, a total of261subjects without any symptom and history of pancreatic diseases underwent whole-body18F-FDG PET-CT in our centre. The maximum SUV (SUVmax) was measured in pancreatic three portions, maximal uptake area of the whole pancreas (whole pancreas), liver, spleen, and muscle, respectively. The subjects were divided into different groups according to gender, age, BMI, history of diabetes mellitus, level of fasting blood glucose, dose of18F-FDG and time interval between injection of18F-FDG and PET-CT scan. SPSS13.0software was used for independent-sample T test, One-Way ANOVA, General linear model and Pearson correlation. A P value<0.05was considered to be statistically significant.Results SUVmax in normal pancreatic head, body, tail and whole pancreas were1.700±0.3675,1.729±0.4105,1.482±0.3983and1.872±0.3950, respectively. SUVmax in normal pancreatic body, tail and whole pancreas were higher in males than in females, and also higher in obesity group than in normal group. The differences were statistically significant (P<0.05). SUVmax in normal pancreatic tail were higher in young and middle-aged groups than in elder group, and also higher in the group of over90min interval than in the group of less than90min interval. The differences were statistically significant (P<0.05). Multivariate analysis revealed that gender and time interval were main influencing factors of SUVmax in normal pancreatic tail. The presence or absence of diabetes mellitus, fasting blood glucose levels (<6.0mmol/1,6.0～6.9mmol/1,7.0～7.9mmol/1and>8.0mmol/1) and injection doses had no statistically significant differences of the impact upon SUVmax of different anatomic portions and maximal uptake area of the whole pancreas (P>0.05). There was slight correlation between SUVmax in pancreas and in spleen or muscle. There was no correlation between SUVmax in pancreas and in liver.Conclusion We preliminary obtain SUVmax in normal pancreatic head, body, tail and maximal uptake area of the whole pancreas in healthy people for cancer screening. Gender, age, BMI, and especially, time interval between injection of18F-FDG and PET-CT scan, have significant influence on18F-FDG uptake in normal pancreas. Whether has diabetes mellitus or not, fasting blood glucose level lower than8.0mmol/1have no effect on18F-FDG uptake in normal pancreas. Objective We studied18F-FDG PET-CT images retrospectively, to optimize the diagnostic cut-off value of the SUVmax and evaluate the value of SUVmax in differentiation of benign from malignant pancreatic masses.Methods18F-FDG PET-CT scan was performed on123patients with pancreatic malignant or benign masses.①Eighty-three patients presented with suspected pancreatic cancer. The final diagnosis in the study was based on pathological or cytological diagnosis (n=33) and clinical and/or diagnostic imaging follow-up for at least6months (n=50). Sixty-eight of them had delayed scans.②Forty patients (44lesions) with pancreatic benign lesions were followed up with clinical and/or diagnostic imaging methods for at least6months. Twenty-six of them (30lesions) had delayed scans. ROC analysis and Student’s t test were used to analyze the data.Results The SUVmax of pancreatic cancer ranged from1.2to12.0(median4.1) in initial scan (n=83); and from1.4to12.8(median4.6) in delayed scan (n=68). The SUVmax of benign lesions ranged from0.7to5.7(median1.5) in initial scan (n=44) and from0.5to6.7(median2.25) in delayed scan (n=30). The difference between benign and malignant groups was statistically significant (t=8.935, P=0.000). The area under the curve of ROC was0.8909in initial scan. With a cut-off value of2.4for the SUVmax, PET-CT had a maximum Youden’s index of0.6670. The sensitivity, specificity and accuracy for diagnosing pancreatic malignant lesions were93.98%,72.73%, and86.61%, respectively. The area under the curve of ROC was0.8596in delayed scan. With a cut-off value of3.2for the SUVmax, PET-CT had a maximum Youden’s index of0.6451. The sensitivity, specificity and accuracy for diagnosing pancreatic malignant lesions were91.18%,73.33%, and85.71%, respectively. The difference of SUVmax between initial and delayed scan of malignant pancreatic lesions was statistically significant (t=-8.834,P=0.000). Those of benign pancreatic lesions showed no statistical significance (t=-1.069, P=0.294).Conclusion The optimal cut-off values of SUVmax in the initial and delayed18F-FDG PET-CT scan for differentiation of benign from malignant pancreatic masses are2.4and3.2, respectively. Objective To compare the accuracy of18F-FDG PET-CT and contrast-enhanced CT (CECT) in T and N staging of pancreatic cancer, to discuss the value of PET-CT in pre-operation M staging of pancreatic cancer, and to analyze the correlation between SUVmax and distant metastasis.Methods From September2006to September2011,60patients with pancreatic cancer underwent18F-FDG PET-CT while59of them underwent CECT. The median time interval between PET-CT and CECT was6days. The final diagnosis in the study was based on pathological or cytological diagnosis (n=30) and clinical and/or diagnostic imaging follow-up for at least6months (n=30). For patients undergoing surgical resection or exploratory procedure, the operational findings were used to assess the accuracy of PET-CT and CECT in T and regional N staging of pancreatic cancer.Results Surgical resection or exploratory procedure was performed in20patients, and peripancreatic organic involvement and/or vascular invasions were found in all the patients. Among them19underwent both PET-CT and CECT. The diagnostic accuracy of CECT and PET-CT for preoperative T staging was94.74%(18/19) and15.00%(3/20), respectively. For the staging of regional lymph nodes, the sensitivity, specificity, and accuracy of CECT were66.67%(8/12),100.00%(8/8), and80.00%(16/20), while of PET-CT were75.00%(9/12),100.00%(8/8) and85.00%(17/20), respectively. There was no significant difference between PET-CT and CT in diagnosing regional lymph node metastases (P=1.00). PET-CT detected hematogenous metastases and distant lymph node metastases in27patients, while conventional imaging detected only in13of them. PET-CT confirmed or excluded hepatic malignancy in6patients when CT was equivocal. Of the60patients, PET-CT detected additional synchronous cancer in2patients. The difference of SUVmax between patients with hematogenous metastasis and without was statistically significant (P=0.026). ROC curve analysis showed that distant metastasis had positive correlation with higher SUVmax. The area under the curve of ROC was0.673. The cut-off value of SUVmax was4.8. But there was no correlation between SUVmax and CA19-9level (r=0.015,P=0.907).Conclusion The accuracy of PET-CT is far lower than that of CECT in T staging for pancreatic cancer. PET-CT is similar to CECT in sensitivity, specificity and accuracy in regional lymph node staging. However, PET-CT is significantly superior in M staging, and can detect synchronous primary cancer. Hematogenous metastasis of pancreatic cancer has positive correlation with higher SUVmax. Objective To preliminary investigate the relationship of18F-FDG uptake with VEGF and MVD in resected specimens of pancreatic cancer. To find out whether18F-FDG uptake can assess tumor angiogenesis in vivo.Methods From May2008to September2011,10patients (6males and4females; median age,61years; age range,41-73years) with pancreatic ductal adenocarcinoma, who had undergone preoperative18F-FDG PET-CT scan without prior chemotherapy or radiotherapy, were enrolled in this study. The PET-CT data was used to calculate tumor maximum SUV (SUVmax). With resected tumor specimens, VEGF and intratumoral MVD expression level were assessed by using immunohistochemical staining. All the patitents were follow-up after having operation.Results In the10patients included in the study,4had carcinoma of pancreatic head and6had carcinoma of pancreatic body or tail. Preoperative PET-CT examination revealed regional lymph node metastases and distant metastases in6patients. The median SUVmax was4.7(range3.3～5.9). The median MVD was67(range49～110). Four patients had negative VEGF expression,3had weak positive expression and3had positive expression. Using median SUVmax4.7as the cut-off value, the10patients were divided into two groups. For patients with SUVmax>4.7,4had regional and distant lymph node metastases while1had synchronous liver metastasis (4/5,80%). Four patients died in progression of pancreatic cancer and1patient was still alive at the end of the study (survival time:6months). The1-year survival rate was0%. Three patients had positive VEGF expression and2patients had weak positive or negative expression. Four patients had high MVD (>67). For patients with SUVmax<4.7,2had regional lymph node metastases and distant metastases. One of them had regional and distant lymph node metastases, and another one had regional lymph node metastasis and peritoneal metastasis. Three patients died in progression of pancreatic cancer and2patients were still alive at the end of the study (survival time:more than12months). All5patients had negative or weak positive VEGF expression. Four patients had low MVD (<67). One patient had slightly higher MVD of69(median MVD67). In3patients with positive VEGF expression, SUVmax were all greater than4.7. Patients with positive VEGF expression had higher SUVmax than those with negative or weak positive VEGF expression. Five patients had higher MVD than the median value of67and4of them had increased SUVmax (>4.7). There was a tendency of positive correlation between SUVmax and MVD in pancreatic ductal adenocarcinoma.Conclusion In this small sample research of10patients, we find out that in patients with pancreatic ductal adenocarcinoma, higher SUVmax of primary tumor suggests more likely existence of metastases. There is a tendency of positive correlation between SUVmax and VEGF, MVD. The preliminary results suggest that18F-FDG PET-CT may be able to assess the angiogenesis in vivo and offer more valuable information for clinical molecular-targeted therapy.