Lung Adjuvant Cisplatin Evaluation (LACE): A pooled analysis of five randomized clinical trials including 4,584 patients.

7008 Background: Several recent trials have shown a benefit of adjuvant cisplatin-based chemotherapy on overall survival (OS) in patients with non-small cell lung cancer (NSCLC). The aim of the Lung Adjuvant Cisplatin Evaluation (LACE) is to identify treatment options associated with a higher benefit, or groups of patients benefiting more from adjuvant chemotherapy. METHODS Individual patient data were collected and pooled from the five largest trials (ALPI, ANITA, BLT, IALT and JBR10) of cisplatin-based chemotherapy in completely resected patients, conducted after the NSCLC-meta-analysis (BMJ 1995, update ongoing). The interactions between patient subgroups or treatment types and chemotherapy effect on OS were analysed using hazard ratios (HR) and logrank tests stratified by trial. RESULTS With a median follow-up of 5.1 years, the overall HR of death was 0.89 (95% confidence interval [CI]: 0.82-0.96; p<0.005) corresponding to a 5-year absolute benefit of 4.2% with chemotherapy. There was no heterogeneity of chemotherapy effect among trials. The benefit varied with stage (test for trend, p=0.046) with the HR for stage I-A 1.41 [95% CI: 0.96-2.09], stage I-B 0.93 [0.78-1.10], stage II 0.83 [0.73-0.95] and stage III 0.83 [0.73-0.95]. The effect of chemotherapy did not vary significantly (test for interaction, p=0.10) with the associated drugs: vinorelbine (HR=0.80 [0.70-0.91]) etoposide/vinca-alcaloide (0.93 [0.80-1.07]) or other (0.98 [0.84-1.14]). There was no interaction between chemotherapy and sex, age, planned radiotherapy or planned total dose of cisplatin. CONCLUSIONS Adjuvant cisplatin-based chemotherapy improves survival in patients with NSCLC. This benefit depends on stage and is greatest in patients with stages II and III. Our analysis suggests that platinum-based adjuvant chemotherapy may not benefit stage I-A patients. Results of disease-free survival will be presented at the meeting. Supported by PHRC and LNLCC [Table: see text].