Restricted mean survival time in advanced non-small cell lung cancer treated with immune checkpoint inhibitors.
Autori
Lorenzo Di Spazio, Luca Cancanelli, Melania Rivano, Marco Chiumente, Daniele Mengato, Andrea Messori.
Rivista
Eur Rev Med Pharmacol Sci
Topic
Analisi statistiche e metanalisi
Impact Factor
3,024
Abstract
Objective: The purpose of this study was to review the effectiveness of immune checkpoint inhibitors (ICIs) in the first-line treatment of advanced non-small cell lung carcinoma with wild-type epidermal grow factor receptor (EGFR) or anaplastic lymphoma kinase.
Materials and methods: After a standard literature search, we identified all randomized studies published on this issue. Our first inclusion criterion was the use of pembrolizumab, nivolumab, atezolizumab or durvalumab in the treatment arm versus chemotherapy in the control arm. The second criterion was the availability of information on overall survival at 2 years. The restricted mean survival time (RMST) was used to analyze the survival curves and rank the treatments.
Results: From the eligible studies, we selected 5 randomized trials that met our inclusion criteria. These trials studied a total of 11 cohorts of patients in whom the treatment arm received ICI as monotherapy (n=3) or in combination with either chemotherapy (n=2) or other monoclonal antibodies (n=1). All the control groups (n=5) received chemotherapy. Pembrolizumab (alone or in combination) showed improvement in overall survival compared with controls, but with borderline statistical significance. Nivolumab, atezolizumab and durvalumab failed to demonstrate any survival advantage. Overall, the RMSTs provided more conservative results than those previously reported using the hazard ratio. In comparing the values of RMST across treatments, pembrolizumab combined with chemotherapy ranked first.
Conclusions: Our results summarized the efficacy of these treatments and showed that only pembrolizumab can have a role as the first-line treatment of NSCLC. These findings are at variance with those previously reported using the hazard ratio as the outcome measure.
Link PubMed del paper
https://pubmed.ncbi.nlm.nih.gov/33660798/