Study design characteristics.  There were a total

of 804 patients in the selected studies and the age ranged from 23 to 94 years. In 15 studies, the sex distribution was described: 385 patients were male and 321 patients were female. In all studies, imaging data were presented about the identification of patients; the reference standard was histopathologic analysis and clinical follow-up. Of all 16 studies, 10 studies28,30–34,36–38,41 enrolled patients 5-Fluoracil prospectively, five studies27,29,35,39,40 enrolled patients retrospectively, and one study26 was unknown. Eight studies27,28,30–35 enrolled patients in a consecutive manner; the others26,29,36–41 were not in a consecutive manner or unknown. There were nine studies26,27,29,31,33–37,40 in which the MRI or PET/CT reviewer was blinded to other test results and clinical data. For DWI, all of the included studies26,29,30,32,33,35,36 used the 1.5T system.

There were three studies26,30,35 in which the average lesion size was over 30 mm. In the other four studies29,32,33,36 the average lesions size was less than 30 mm. For PET/CT, contrast enhanced PET/CT was used in four studies,27,28,31,40 and noncontrast enhanced PET/CT was used in seven check details studies,27,31,34,37–39,41Table 1 presents the included datasets with the corresponding numbers of patients and reference numbers. A full list of all included articles with all relevant study characteristics and complete examination results is available on request from the authors of this article. Diagnostic accuracy

of 18 F-FDG PET and DWI.  When considering all 16 studies with data on pancreatic malignancy per patient, for PET/CT, the pooled sensitivity was 0.87 (95% CI, 0.82, 0.81) and specificity was 0.83 (95% CI, 0.71, 0.91). Overall, LR+ was 5.84 (95% CI, 4.59, 7.42) and LR− was 0.24 (95% CI, clonidine 0.17, 0.33). For DWI, the pooled sensitivity was 0.85 (95% CI, 0.74, 0.92) and specificity was 0.91 (95% CI, 0.71, 0.98). LR+ was 9.53 (95% CI, 2.41, 37.65) and LR− was 0.17 (95% CI, 0.09, 0.32). SROC curves show the overall very good, but not excellent, diagnostic performance for PET/CT and DWI is shown in Figures 2 and 3, respectively. Subgroup analysis and meta-regression analysis. I2 is an index for heterogeneity: I2 = [Q − (k − 1)]/Q × 100%, where Q is the χ2 value of heterogeneity, and k is the number of studies included. Along with P < 0.05 for heterogeneity, I2 > 50% further indicates heterogeneity between studies. The heterogeneity in the sensitivity test and specificity test was highly significant (P < 0.05 and I2 > 50%), confirming that there was strong evidence of between-study heterogeneity both for PET/CT and DWI (Table 2). Therefore, a random effect model was used for the primary meta-analysis to obtain a summary estimate for sensitivity and specificity with 95% CI. To explore the possible source of heterogeneity, subgroup analyses were applied (Table 2).