RESUMEN
BACKGROUND/AIM: P21 is a cyclin-dependent kinase inhibitor regulating the cell cycle as a tumor suppressor. Using a p21 immunohistochemistry (IHC) assay, we compared tumor p21 levels with conventional clinico-pathological criteria in primary pancreatic endocrine tumor subsets with and without liver metastases. MATERIALS AND METHODS: Sections from tissue microarray (TMA) including 13 archival metastatic primary and 18 non-metastatic primary pancreatic endocrine carcinomas/tumors (MP-PECAs/NMP-PETs) were stained with a monoclonal anti-p21WAFI,CIP primary antibody. Tumor p21 IHCs were scored as the sum of intensity (0-3) and proportion scores (0-5) (Total Allred score: 0-8), and as p21% labelling index in the tumor. ROC curve analysis was used for most optimal p21 score cut-off (4 or >) and Fisher's exact test was used to compare the association among tumor p21 scores, conventional prognostic criteria, and liver metastases. RESULTS: For PET/PECA patients, mean ages were 55.6 years (27-73) and 49.3 years (28-71), M/F ratios were 7/11 and 7/6. Mean p21 labelling index (%) for MP- PECAs was 24% (range=3-63%) vs. 9% for NMP-PETs (range=1-25%) (p=0.022). The mean p21 index in MP-PECAs was significantly higher (24%) as compared to PIs (7%) (p=0.0047). Using a p21 Allred score of ≥4, high p21 IHC score had strong association with the presence of liver metastases (p-value <0.001). High tumor p21 IHC score had a 93% sensitivity, 68% specificity, 78% predictive accuracy, 66% positive, and 94% negative predictive values. CONCLUSION: In patients with primary PETs, p21 IHC is superior to conventional criteria in predicting presence or absence of liver metastases.