“…These 2 internal weaknesses increased the risk of interpretation biases when building the 2 by 2 table "a priori." Moreover, in the meta-analysis by Kim et al, 5 10 studies did not report a clear definition of a BD-IPMN, 28,30,31,[34][35][36][37]46,47,49 including possible mixed-type IPMNs in the analysis and, consequently, increasing the rate of malignancy in the study population whereas, in the present study, all BD-IPMNs were clearly defined. Finally, in the present meta-analysis, all risk factors were considered, not only radiological parameters, as reported by Kim et al 5 ; the diagnostic performance of each parameter was evaluated using several classic biostatistical measurements of accuracy (DOR, sensitivity, specificity, AUC, positive and negative predictive values, or positive and negative likelihood ratios) whereas, in the previous metaanalyses, very few biostatistical measurements were adopted to explain the results.…”