The incidence of brain metastases secondary to small cell lung cancer (SCLC) is about 35% and the treatment strategy of brain irradiation with respect to dose and fractionation is controversial. In order to evaluate treatment outcome of brain irradiation in SCLC patients with brain relapse, we retrospectively evaluated all patients treated with brain irradiation in the eastern part of Denmark from 1988 to 1992 (PCI patients excluded). During this 5-year period, 101 evaluable patients were included (44 females, 57 males) (median age 61 years; range, 39-75 years). Forty-four patients, of whom 43 were in extracerebral complete remission (CR), received extended course (EC) brain irradiation (>45 Gy, treatment schedule > 4 weeks). Fifty-seven patients received short course (SC) brain irradiation (<30 Gy, treatment schedule < 1 week). Among the SC treated patients, 14 were in CR, 20 had partial remission or stable disease and 23 had progressive extracerebral disease. The median survival (from diagnosis of brain metastases) in the group receiving irradiation with EC (44 patients) was 160 days (range, 74-2021 days), while the 57 patients treated with SC had a median survival of 88 days (range, 20-948 days) (P = 0.00001, Log-Rank analysis). In a subgroup of 14 patients in extracerebral CR, receiving SC irradiation, the median survival was 83 days (range, 15-948 days). When the latter patients were compared to the 43 patients in CR in the group treated with EC, a statistically significant difference was shown (P = 0.034, Log-Rank analysis). Using Cox-hazard regression analysis with backward elimination, liver metastases and poor performance status were adverse prognostic signs, although the only significant parameters of survival were gender (female vs. male, relative risk of dying 1 and 1.52, P = 0.05) and schedule of brain irradiation (extended course vs. short course, relative risk of dying, 0.36 and 1, P < 0.001). Extended course irradiation of brain relapse secondary to SCLC seems in general to be of limited value, although a significant prolonged survival at approximately 7 weeks, was obtained. The prolongation of survival does not seem worthwhile considering the length of treatment time (5-6 weeks) compared to SC treatment (1 week). However, the data do not permit evaluation of the quality of life of the patients. This retrospective evaluation suggests the need for randomized trials with carefully planned quality-of-life assessments.