How reliable is the Ki‐67 cytological index in grading pancreatic neuroendocrine tumors? A meta‐analysis

Abstract

To investigate the accuracy of the cytological Ki-67 index in distinguishing intermediate and high-grade (G2 + G3) from low-grade (G1) pancreatic neuroendocrine tumors (PNETs). Two investigators independently searched databases to identify eligible studies using the following term: ('Ki-67') AND ('pancreatic endocrine tumor' OR 'pancreatic neuroendocrine tumor' OR 'pancreatic endocrine tumour' OR 'pancreatic neuroendocrine tumour' OR 'pancreatic endocrine tumors' OR 'pancreatic neuroendocrine tumors' OR 'pancreatic endocrine tumours' OR 'pancreatic neuroendocrine tumours'), and meta-analysis was performed to calculate the pooled sensitivity, specificity, positive (PLR) and negative likelihood ratio (NLR), and diagnostic odds ratio (DOR). A total of 263 lesions from 13 studies were included in the study. The pooled sensitivity and specificity of Ki-67 (cut-off value: 2%) in the differential diagnosis of G2 + G3 from G1 PNETs were 64% and 87%, respectively. The pooled PLR, NLR and DOR were 3.96, 0.42 and 11.21, respectively. The area under the summary receiver operating characteristic curve (AUROC) was 0.8397. While the cut-off value of Ki-67 index was set as 5%, the sensitivity and specificity were increased up to 69% and 93%, respectively, and the AUROC was increased to 0.955. The cytological Ki-67 index is very useful in distinguishing intermediate and high-grade from low-grade PNETs, and a cut-off value of 5% had a better predictive value compared with that of 2%.