Abstract

Background The optimum dose of prophylactic cranial irradiation (PCI) for limited-stage small-cell lung cancer (SCLC) is unknown. A meta-analysis suggested that the incidence of brain metastases might be reduced with higher PCI doses. This randomised clinical trial compared the effect of standard versus higher PCI doses on the incidence of brain metastases. Methods Between September, 1999, and December, 2005, 720 patients with limited-stage SCLC in complete remission after chemotherapy and thoracic radiotherapy from 157 centres in 22 countries were randomly assigned to a standard (n=360, 25 Gy in 10 daily fractions of 2·5 Gy) or higher PCI total dose (n=360, 36 Gy) delivered using either conventional (18 daily fractions of 2 Gy) or accelerated hyperfractionated (24 fractions in 16 days with two daily sessions of 1·5 Gy separated by a minimum interval of 6 h) radiotherapy. All of the treatment schedules excluded weekends. Randomisation was stratified according to medical centre, age (≤60 and >60 years), and interval between the start of induction treatment and the date of randomisation (≤90, 91–180, and >180 days). Eligible patients were randomised blindly by the data centre of the Institut Gustave Roussy (PCI99-01 and IFCT) using minimisation, and by the data centres of EORTC (EORTC ROG and LG) and RTOG (for CALGB, ECOG, RTOG, and SWOG), both using block stratification. The primary endpoint was the incidence of brain metastases at 2 years. Analysis was by intention-to-treat. This study is registered with ClinicalTrials.gov number NCT00005062. Findings Five patients in the standard-dose group and four in the higher-dose group did not receive PCI; nonetheless, all randomised patients were included in the effectiveness anlysis. After a median follow-up of 39 months (range 0–89 months), 145 patients had brain metastases; 82 in the standard-dose group and 63 in the higher-dose group. There was no significant difference in the 2-year incidence of brain metastases between the standard PCI dose group and the higher-dose group, at 29% (95% CI 24–35) and 23% (18–29), respectively (hazard ratio [HR] 0·80 [95% CI 0·57–1·11], p=0·18). 226 patients in the standard-dose group and 252 in the higher-dose group died; 2-year overall survival was 42% (95% CI 37–48) in the standard-dose group and 37% (32–42) in the higher-dose group (HR 1·20 [1·00–1·44]; p=0·05). The lower overall survival in the higher-dose group is probably due to increased cancer-related mortality: 189 patients in the standard group versus 218 in the higher-dose group died of progressive disease. Five serious adverse events occurred in the standard-dose group versus zero in the higher-dose group. The most common acute toxic events were fatigue (106 [30%] patients in the standard-dose group vs 121 [34%] in the higher-dose group), headache (85 [24%] vs 99 [28%]), and nausea or vomiting (80 [23%] vs 101 [28%]). Interpretation No significant reduction in the total incidence of brain metastases was observed after higher-dose PCI, but there was a significant increase in mortality. PCI at 25 Gy should remain the standard of care in limited-stage SCLC. Funding Institut Gustave-Roussy, Association pour la Recherche sur le Cancer (2001), Programme Hospitalier de Recherche Clinique (2007). The European Organisation for Research and Treatment of Cancer (EORTC) contribution to this trial was supported by grants 5U10 CA11488-30 through 5U10 CA011488-38 from the US National Cancer Institute.