How much can patients, with low Karnofsky performance status and advanced non-small cell lung cancer, benefit from cisplatin-based chemotherapy? A retrospective analysis of a phase III randomised trial

Abstract

19516 Background: Performance status (PS) is a strong independent prognostic factor for survival in patients with advanced non- small cell lung cancer and administration of platinum-based chemotherapy (CT) in patients with poor PS is associated to more severe toxicity. Our purpose was to determine whether poor PS patients (ie those registered with a performance index of 60–70 on the Karnofsky scale) could benefit from CT in terms of clinical improvement defined as reaching, during CT, a PS of at least 80 on the same scale (ECOG PS 0 or 1). Methods: Retrospective analysis of a randomised trial performed in advanced NSCLC where 485 patients received three courses of GIP (gemcitabine + ifosfamide + cisplatin) induction CT, in order to determine the potential clinical benefit of conventional cisplatin-based CT in patients with poor PS defined as 60–70 on the Karnofsky scale. Results: 387 (80%) patients had good PS (Karnofsky 80–100) and 98 (20%) poor PS. Response rates were respectively 38 and 28% (p=0.06), no independent predictive value for PS was found using response to CT as endpoint. Clinical improvement was observed in 25% (95% CI : 15%-38%) of the poor PS patients (respectively 38%, 20% and 14% in case of response, no change and progression, p=0.05). Nine additional patients improved from 60 to 70 during induction CT. Improvement occurred earlier in responders (rates of patients with improvement were 12% after 1 course, 31% after 2 courses and 38% after 3 courses compared to 3%, 5% and 17% in the non responders). Survival of patients with poor PS was significantly worse but survival of responders was similar, whatever initial poor or good PS (logrank test after landmark, p=0.47, HR=1.18, 95%CI: 0.76–1.84 for poor responders). If non fatal toxicity was relatively similar as well as dose intensity delivered after 3 CT courses, there were more toxic deaths (including vascular and cardiac fatalities) in poor PS patients (2.1% versus 9.2%; p=0.002). Conclusions: Although toxicity is increased, combination chemotherapy is associated with clinical improvement in a non negligible rate of patients with poor PS in case of advanced NSCLC. No significant financial relationships to disclose.