The value of cell kinetic measurements in head and neck tumors in predicting which patients will benefit from accelerated fractionation radiotherapy regimens is being tested in a multicenter European trial (EORTC trial 22851). This paper reports on the first analysis of the correlation of kinetics with outcome in this trial. A proportion of patients in both the accelerated arm (72 Gy in 5 weeks, 1.6Gy per fraction, 45 fractions) and the conventional arm (70-72 Gy in 7-8 weeks, 1.8-2.0 Gy per fraction, 35-40 fractions) were given an i.v. injection of 100 mg/m2 IUdR (iododeoxyuridine) before treatment, and a tumor biopsy was taken several hours later. The potential doubling time of the tumor (Tpot) was obtained from a flow cytometric analysis of tumor cell nuclei using an anti-IUdR antibody. From a total of 260 patients entered in the trial, 53 have undergone kinetic analysis. Adequate IUdR labeling was seen in 47 patients (88.7%), from which the mean value for Tpot was found to be 4.5 +/- 2.5 days (+/- S.D.). Of the IUdR labeled patients, 30 have now been followed up for at least 1 year, 17 with conventional and 13 with accelerated radiotherapy. These patients were split into those with fast and those with slowly growing tumors, the dividing line being the median Tpot value of 4.6 days. After conventional 7-week radiotherapy, 2 of 6 patients with "fast" growing tumors obtained local control compared with 8 of 11 with "slow" growing tumors. A small difference in local control was seen been fast and slow tumors in the accelerated arm (5/9 vs. 3/4). These preliminary data support the hypothesis that patients with fast growing tumors do poorly with conventional radiotherapy and that pretreatment kinetic measurements can select patients at risk. The predictive power of the method must await the final analysis of trial results.