To determine the impact of research findings and evolving technology on the patterns of radiotherapy practice for patients with carcinoma of the uterine cervix. Sixty-two radiation therapy facilities participated in the study after having been selected from a random sample, proportionally stratified according to practice type, of all United States facilities. Each facility submitted a list of patients treated during 1992-1994 with radiation for squamous carcinoma of the cervix. Cases for review were randomly selected from each institution after excluding those of patients who had distant metastases or initial hysterectomy. A total of 471 patients' records were reviewed in the treating institutions to obtain information about patients' characteristics, diagnostic evaluation, tumor extent, treatment approach, and radiotherapy techniques. Of the 61 facilities that treated eligible cases of intact cervical cancer during the 3-year survey period, 35 (57%) treated fewer than three eligible patients per year. Thirty-four (83%) of 41 non-academic facilities vs. 1 (5%) of 20 academic facilities treated fewer than three patients per year. FIGO stages were I, II, III, and IV in 32%, 40%, 24%, and 3% of patients, respectively. Computed tomography (CT) was the most common method of lymph node evaluation, but surgical evaluation, which was performed in 76 (16%) patients, had increased from previous surveys. Fields were designed using a dedicated simulator in 95% of patients; a dedicated CT unit was used for treatment planning in 119 (30%) cases. External beam irradiation was most often given using a four-field technique at 180 cGy per day on a 10-20 MV linear accelerator. The average daily fraction size had decreased from previous surveys, and 13% of patients were treated with daily doses of 170 cGy or less. Most patients were treated with a combination of external beam and low dose-rate (LDR) intracavitary irradiation. Of 425 patients who had treatment with curative intent that included brachytherapy, 362 (85%) had LDR brachytherapy, 45 (11%) had high dose-rate (HDR) brachytherapy, 3 had a combination of HDR and LDR, and 15 had incomplete information about the brachytherapy dose-rate. Forty-six (23%) of 197 patients with Stages I-IIA disease were treated with radiation followed by extrafascial hysterectomy. Of 111 patients treated with curative intent for Stage III-IV disease, 72 (65%) had a combination of external beam and intracavitary radiation therapy, 22 (20%) had external beam plus interstitial brachytherapy, and 17 (15%) were treated with external beam irradiation only. For patients who completed treatment with curative intent and did not have adjuvant hysterectomy or HDR brachytherapy, the median total dose at point A was 82.5 Gy. For all patients who completed treatment with radiation alone, the median total duration of treatment was 63 days; more than 70 days were taken to complete treatment in 33% of cases. Twenty-nine percent of patients received chemotherapy, usually concurrent with their radiation therapy. Only 27% of these patients were on investigational protocols. Greater participation in well-designed prospective trials might help clinicians address important clinical questions and reduce current inconsistencies in the use of adjuvant treatments. Radiation oncologists should take steps to avoid unnecessary treatment protraction and to improve patient compliance. Future studies will be needed to determine whether the small number of cases being treated in most nonacademic facilities will influence the outcome for patients with invasive cervical carcinoma.