Background
Thrombospondin-2 (TSP-2) has been reported as an early diagnostic marker for pancreatic ductal adenocarcinoma (PDAC) in Caucasian populations. This study was designed to validateTSP-2 as a diagnostic marker in a large Taiwan cohort and to investigate the association of TSP-2 with the clinical outcomes of PDAC patients.Methods
The serum TSP-2 levels in 263 PDAC patients and 230 high-risk individuals (HRIs) were measured via an enzyme-linked immunosorbent assay. The sensitivity, specificity, and accuracy of TSP-2 as a diagnostic marker to discriminating PDAC patients from HRIs and correlations between TSP-2 levels and prognosis of PDAC patients were analyzed.Results
Serum TSP-2 levels were significantly higher in patients with PDAC (44.90 ± 40.70 ng/ml) than in the HRIs (17.52 ± 6.23 ng/ml). At a level of ≥ 29.8 ng/ml, TSP-2 exhibited 100% specificity, 55.9% sensitivity, 100% positive predictive value (PPV), and 66.5% negative predictive value (NPV) for discriminating PDAC patients from HRIs. The Cox regression analysis showed that higher serum TSP-2 levels were significantly associated with poor outcomes in PDAC patients (hazard ratio = 1.54, 95% confidence interval = 1.143-2.086, P = 0.005). Combining the carbohydrate antigen 19-9 (CA19-9) (cutoff value of 62.0 U/ml) and TSP-2 (cutoff value of 29.8 ng/ml) levels yielded 98.7% specificity, 90.5% sensitivity, 98.8% PPV, and 90.1% NPV for discriminating patients with PDAC from HRIs.Conclusions
TSP-2 is a highly specific diagnostic marker and an independent prognostic marker in patients with PDAC. A combined biomarker panel, including TSP-2 and CA19-9, may facilitate future PDAC screening.