Radiotherapy in laryngeal carcinoma: can a panel of 13 markers predict response?

OBJECTIVES/HYPOTHESIS: To find biomarkers associated with response to radiotherapy in laryngeal cancer that can be used together with clinical parameters to improve outcome prediction. METHODS: In this study, 26 patients irradiated for laryngeal carcinomas with a local recurrence within two years (cases) and 33 patients without recurrence (controls) were included. All pretreatment biopsies were arrayed onto a tissue array. Immunohistochemistry was performed for 13 biomarkers that were selected from the literature as potential predictors for radioresponse in head and neck (HN) cancer: Bcl-2, Bcl-xL, p16, p21, p27, p53, cyclin D1, HIF-1alpha, CA9, COX-2, EGFR, ki-67, and pRB. RESULTS: Univariate logistic regression models showed borderline statistically significant increased relative risks, with positivity for CA9, COX-2, and p53. Goeman's global testing revealed an overall association between outcome and the 13 markers together with clinical variables. The most important markers were CA9 and COX-2. CONCLUSIONS: In laryngeal carcinoma, hypoxia and COX-2 overexpression provide a stronger contribution to an increased risk of local recurrence after radiotherapy compared with the well-known candidate markers p53, Bcl-2, and cyclin D1. However, no robust expression profile for the prediction of radioresistance was found