The identification of markers predicting the response to therapy is of the utmost importance in oncology. Several authors have suggested that increased levels of glutathione (GSH) and glutathione S-transferase (GST) activity might be meaningful predictors of poor responsiveness to chemotherapy in several human cancers, but the biological assays have not been standardised and published studies show conflicting evidence. The aim of the present study was to select a validated panel of tests to assess the GST/GSH system in a clinical setting. Matched blood and tissue samples (normal and malignant) from 52 cancer patients with either non-small cell lung cancer (NSCLC) or head and neck squamous cell carcinoma (SCCHN) were investigated. GSH levels and GST activity were higher in cancer tissues than in matched normal tissues in both malignancies. The difference was statistically significant in NSCLC (P=0.0004 and P=0.0002, for GSH and GST, respectively) and borderline in SCCHN (P=0.03 and P=0.02, for GSH and GST, respectively). Moreover a strong correlation was found between the GSH level in whole blood and GST activity in cancer tissue in both malignancies (P=0.003, r=0.53 in NSCLC, P<0.0001, r=0.89 in SCCHN). In conclusion, reliable and robust methods for routine use in tissue extracts and in whole blood have been validated. Our finding regarding the GSH level in blood indicates that circulating GSH could have a clinical relevance as a surrogate marker of GST activity in tumour tissue.