Summary Patients with T2 grade 3 and T3 bladder cancer were randomised to be treated with radiation alone (NO MISO) or with radiation and misonidazole (PLUS MISO). Patients in both groups initially received 40 Gy in 2 Gy fractions (5/week). Patients in the NO MISO arm received a further 20 Gy in 2 Gy fractions (5/week). Patients in the PLUS MISO arm received a further 12 Gy in 6 Gy fractions (1/week).MISO was administered orally (3.0 g m-2) and intravesically (1.0 g in 35 ml of solvent) 4 h and 2 h respectively prior to each fraction of 6 Gy.Fifty-eight patients were randomised of whom 53 are evaluable. There is a minimum follow-up of 5 years in the surviving patients. In the NO MISO and PLUS MISO arms, the complete response rate at cystoscopy at 6 months was 63% and 69%, the 5-year survival rate was 41% and 48% and the 5-year local control rate with bladder preservation was 46% and 36% respectively (censored for death from metastases while locally clear). Misonidazole (MISO) has been experimentally shown to sensitise hypoxic tumour cells to irradiation both in vitro and in vivo (Adams, 1978). The sensitiser enhancement ratio of MISO is dependent on the tumour concentration of the drug (Asquith et al., 1974). In clinical studies of fractionated radiation the total amount of MISO which can be administered is limited by the drug's neurotoxicity (peripheral neuropathy) (Dische et al., 1979), and patients have been treated at low tumour concentrations of MISO.In patients with bladder carcinoma, high tumour concentrations of MISO are obtainable after intravesical administration (Awwad et al., 1983). A treatment regimen was devised in which an initial course of conventionally fractionated radiation was followed by two administrations of oral and intravesical MISO plus large fractions of radiation (Abratt, 1982). The regimen was designed with the specific aim of achieving high tumour concentrations of MISO. The aim was to radiosensitise and sterilise clonogenic hypoxic cells persisting after conventionally fractionated radiation.The results of a pilot study of this regimen was apparently better than that in a series of historical controls (Abratt et al., 1983;Abratt et al., 1987 Table I.