Abstract
Improved therapies for individuals with head and neck squamous cell carcinoma (HNSCC) may be developed by identification of appropriate biomarkers. The aim of this study was to evaluate the usefulness of serum midkine measurement as a biomarker for HNSCC. Pretreatment serum midkine concentrations were measured in 103 patients with HNSCC and 116 control individuals by enzyme-linked immunosorbent assay. Midkine expression in tumor tissues from 33 patients with HNSCC who underwent definitive surgical resection without preoperative treatment was examined by immunohistochemistry. The cut-off serum midkine concentrations for predicting the presence of head and neck malignancy and chemosensitivity to induction chemotherapy, as determined using receiver operating characteristic curves, were 482 and 626 pg/mL, respectively. Spearman bivariate correlations showed positive correlations between serum midkine levels and immunohistochemistry staining score (r = 0.612, P < 0.001). The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of serum midkine concentration for detection of HNSCC were 57.3, 85.3, 77.6, 69.2, and 72.1%, respectively. However, for predicting the response to induction chemotherapy, the values were 84.6, 60.9, 71.0, 77.8, and 73.5%, respectively. Serum midkine concentration was identified as an independent prognostic factor by multivariate analysis, using Cox's proportional hazards model (P = 0.027). Overexpression of serum midkine yielded a relative risk of death of 3.77, with 95% confidence limits ranging from 1.15 to 17.0. Serum midkine levels in patients with HNSCC were associated with malignancy, chemosensitivity, and prognosis. Serum midkine may be a useful, minimally invasive biomarker for early detection, therapeutic decision-making, and predicting prognosis.