A single line source brachytherapy (BT) technique has been developed at Clatterbridge to boost the dose to the primary tumour after whole pelvis external beam radiotherapy (EBRT) for the radical treatment of carcinoma of the cervix. 226 patients with invasive carcinoma of the uterine cervix were treated with radiotherapy alone using this technique (median age 57 years; range 25-87 years). 49 patients had Stage IB disease, 97 had Stage II, 73 had Stage III and seven patients had biopsy confirmed Stage IVA disease. Patients with low bulk disease were given 40-42.5 Gy in 20 fractions while those with bulky disease received 45 Gy in 20 fractions or 50 Gy in 25 fractions. On completion of EBRT, 186 patients (82.3%) proceeded to intracavitary BT using a linear arrangement of sources with the Selectron (Nucletron) remote afterloading unit. Most of the patients (137/226, 60.6%) received a single insertion of 20 Gy to point "A", at a preferred dose rate within the range 0.95-1.05 Gy h-1. In another 30 patients (13.3%), BT was possible at a later date after further tumour regression. Only 10 patients (4.4%) did not receive BT as part of their treatment. The 5 year actuarial cause-specific survival rate was 79% in Stage I disease, 61% in Stage II, 31% in Stage III and 71% in the small number of patients with Stage IVA disease. The 5 year pelvic control rates were 88% for Stage I, 69% for Stage II, 45% for Stage III and 71% for Stage IVA. Significant prognostic variables for survival and local pelvic control on univariate analysis included disease stage, patient age, tumour bulk, nodal status, anaemia, renal failure and overall treatment time. Tumour grade was a significant prognostic variable for survival but not for local tumour control. The extent of parametrial involvement was a significant prognostic variable for survival and local control for Stage IIB but not for Stage IIIB. There was a statistically significant decrease in survival and local tumour control for patients receiving > or = 70 Gy to point "A", or > or = 55 Gy to point "B". On multivariate analysis, the independent prognostic variables for survival and local control were disease stage, overall treatment time and renal failure. Patient age was also an independent prognostic variable for survival while nodal status was an independent prognostic variable for local control. A high proportion of the patients had adverse prognostic features resulting in a very high actuarial risk of distant metastases of 38.1% at 5 years (68.8% for Stage III patients). The overall treatment time was significantly longer in Stage III patients compared with Stage I and Stage II patients. The actuarial rate of Grade 2 late radiation morbidity was 2.7% and 4.3% for the urinary tract and bowel respectively while that of Grade 3 morbidity was only 0.6% and 1.4%, respectively. Good local control can be achieved for patients with nonbulky tumours using relatively low biological doses while minimizing the risk of late treatment related toxicity. Several changes in treatment policy have been made in an attempt to improve local tumour control and possibly survival, particularly for Stage III patients and patients with bulky disease.