It has been suggested that alterations involving the p53 gene may influence tumour response to radiotherapy. If this were so, then p53 overexpression (which is usually associated with p53 mutation and readily detectable in routine diagnostic pathology) may help determine the most appropriate form of cancer therapy. p53 expression was assessed in 90 formalin-fixed paraffin-embedded laryngeal carcinomas that were subsequently treated with radiotherapy. The polyclonal antibody DO1 (1 in 50 dilution) was used, together with an avidin-biotin immunoperoxidase technique, but in the absence of any additional antigen retrieval techniques. p53 expression was assessed and correlated with various clinicopathological parameters. Using Chi square analysis, no significant difference between p53 positive and p53 negative lesions was found for response to radiotherapy, as measured by survival and recurrence rates. Furthermore, no correlation with p53 expression was found for tumour size, nodal metastasis, sex, age, alcohol intake, tobacco habit and histological grade. This absence of correlation may in part be explained by discrepancies between immunohistochemical detection of p53 and p53 mutation, although the lack of predictive response to radiotherapy mimics that recently found for irradiated head and neck cancer cell lines.