Between 1959 and 1980, 590 evaluable patients received a full course of radiation therapy as primary treatment of invasive epidermoid carcinoma of the cervix. At follow-up visits conducted between 1 and 3 months after the completion of radiation, 72 patients had physical or biopsy findings indicative of persistent disease (PER), whereas an additional 36 had findings suggestive of tumor (SUS). The remaining 482 demonstrated a complete response (CR) by 1 month after the administration of treatment. The lower the stage, the more likely was a CR by 3 months (Stage I, 94.4%; Stage II, 86.2%; Stage III, 62.7%). A higher proportion of CR was obtained in patients with tumor of less than 5 cm in diameter than in those with tumors greater than 5 cm in diameter (82.4% versus 76.3%, respectively; 0.1 greater than P greater than 0.05). CR patients enjoyed a marked survival advantage over those with PER, whereas women with SUS demonstrated intermediate 5-year survival (76.0%, 41.5%, and 7.4%, respectively; P less than 0.0001). The same survival gradient of CR greater than SUS greater than PER was demonstrated when tumors of each stage were analyzed separately. There was no difference in survival between patients with PER or SUS at 1 month and those with PER or SUS at 3 months. The proportion of recurrent tumors diagnosed as exclusively distant metastases decreased from CR to SUS to PER. Analysis of dosimetry data suggests that the likelihood of CR was not a function of treatment variability. It was concluded that tumors that do not regress promptly are likely to recur, most with distant metastases. Such poor-prognosis patients should be targeted for early adjuvant or salvage therapy.