The possibility to perform flow cytometry was examined in a series of 167 patients with primary untreated head and neck carcinoma referred to our Institution from February 1989 to January 1992. In all cases, flow cytometry was carried out on frozen tumour samples. The Cox model was used including age, tumour size, nodal status on clinical assessment, topography, treatment, malignancy grade, S phase fraction and ploidy as independent variables and overall survival as dependent variable. In this study, ploidy could be assessed in only 73% of cases and S phase fraction and G2M in 65% of the population studied. No correlation could be evidenced between ploidy or SPF with other clinical, pathologic characteristics or clinical outcome. We conclude that flow cytometry should remain a research tool until the method has proved to be relevant in clinical routine, and until the yield of the technique can be improved.