Objectives: A prior study suggested serum heat shock protein 27 (HSP27) as a potential marker for pancreatic carcinoma, but its accuracy in differentiating cancer from chronic pancreatitis was not evaluated. We aimed to analyze HSP27 levels in pancreatic carcinoma, chronic pancreatitis, and healthy subjects and assess its diagnostic efficacy.
Methods: Pretreatment serums from 58 pancreatic carcinoma, 44 chronic pancreatitis, and 102 control subjects were collected. Serum HSP27 and carbohydrate antigen 19-9 (CA19-9) levels were analyzed using an enzyme-linked immunosorbent assay and radioimmunoassay, respectively.
Results: Heat shock protein 27 levels were significantly higher in cancer and pancreatitis compared with control (P < 0.001 for both), but no significant difference was noted between cancer and pancreatitis (P = 0.978). By logistic regression, HSP27 was a significant predictor of differentiation between cancer and control (P < 0.0001) but not between cancer and pancreatitis (P = 0.885). At a cutoff of 1650 ng/L, the sensitivity and specificity for differentiating cancer from healthy control were 62.1% and 95.1%, respectively. Receiver operating characteristic analyses showed a greater area under curve for CA19-9 compared with HSP27 in differentiating between cancer and control (0.92 and 0.84, respectively, P = 0.014).
Conclusions: Serum HSP27 is increased in both chronic pancreatitis and pancreatic carcinoma. It should not be recommended as a diagnostic marker for pancreatic carcinoma.