Using a microarray technique, we found decorin to be underexpressed, but osteopontin (OPN) to be overexpressed, in esophageal squamous cell carcinoma (ESCC). This study aims to examine whether plasma decorin and OPN plus personal substances use (tobacco, alcohol and areca) can serve as suitable clinical markers to predict the presence of ESCC. In total, 570 archived plasma specimens (275 patients and 295 controls) were collected from 2 medical centers in Taiwan between 2000 and 2008. Decorin and OPN protein levels were measured by ELISA. Means and standard deviation of plasma decorin were 5.6 + or - 3.6 ng/ml in case patients, which were significantly lower than those in controls (7.8 + or - 3.1, p < 0.0001). Plasma OPN levels in case patients were not significantly different from controls (p = 0.33). When compared to subjects with the lowest quartile of plasma decorin, those with the highest quartile one had a significantly lower risk to have ESCC (Adjusted OR = 0.03, p < 0.001). Receiver operator characteristic (ROC) analysis was performed for the combination of plasma decorin and 3 substances use (smoke, alcohol and areca) for the patients compared with the controls. The area under the curve was 88.6% and the optimal cut-point of ROC curve (any 3 factors) had 73.5% sensitivity and 90.2% specificity with approximately 82% of corrected classification. Plasma decorin, but not OPN, is a potential clinical marker for the detection of ESCC. When plasma decorin plus the use of the 3 substances are combined, this factor cluster could be used to detect the presence of ESCC.