Summary Estimation of tumour proliferation may allow the design of individualised radiotherapy schedules to optimise response. This prospective study correlates the tumour proliferation rate of cervical carcinoma with response to conventional radiotherapy. The potential tumour cell doubling rate (Tpot) was estimated following flash labelling of the tumours in vivo using the DNA precursor, bromodeoxyuridine (BrdUrd); samples were analysed by flow cytometry. Tumour ploidy, DNA index and mitotic count were also assessed as was histological grade and type. Multiple biopsies from each tumour were obtained from 121 women. The median Tpot was 4.0 days, median S-phase duration 12.8 h and median adjusted labelling index 9.8%. Higher BrdUrd labelling was seen in patients who developed pelvic tumour recurrence following radiotherapy. This was the only biological/histological parameter with univariate and multivariate significance in relation to locoregional recurrence (P = 0.006 and P = 0.034 respectively). This study represents the first assessment of Tpot in relation to long-term response of cervical tumours treated by radiotherapy treatment. The association of high BrdUrd labelling and poor pelvic disease-free survival indicates the need for further research into the potential of radiotherapy schedule alteration to reflect tumour proliferation. The predictive value may be enhanced by combination with other biological parameters.Keywords: cervical carcinoma; radiotherapy; proliferation rate; bromodeoxyuridine Local recurrence can be a significant problem following radiotherapy for carcinoma of the cervix. This is particularly true of more advanced stage tumours. Local recurrence is seen in only 10% of stage lb patients, but in stage IIIb disease, pelvic recurrence rate is greater than 40% (Davidson et al., 1989). Both clinical and laboratory data suggest repopulation during treatment may be an important factor leading to failure to achieve local control of the tumour (Trott and Kummermehr, 1985). In order to reduce repopulation during treatment, radiotherapy can be given over a shorter time period, often two or three times a day using a fraction size of less than 2 Gy. It is unclear whether all patients would benefit from accelerated radiotherapy schedules. Initial results from a randomised study of radiotherapy schedule alteration in head and neck cancer (Saunders et al., 1991) suggest that local control is improved in patients receiving accelerated hyperfractionated radiotherapy in rapidly proliferating tumours only (Begg et al., 1992). Measurement of tumour proliferation rate may, therefore, help to select patients most likely to benefit from new or accelerated schedules for radiotherapy.There is inconclusive evidence regarding the relationship between measured tumour proliferation parameters and prognosis in cervical carcinoma (Dixon et al., 1977; Strang et al., 1987a,b;Naus and Zimmerman, 1991;Cole et al., 1992;Zanetta et al., 1992;Tsang et al., 1995). The techniques used in the measurement of proliferation in the...