摘要 : Immune checkpoint inhibitors (ICIs), either alone or combined with chemotherapy, represent the cornerstone of the treatment of advanced non-small cell lung cancer (NSCLC) without targetable gene alterations. Programmed death ligan... 展开 Immune checkpoint inhibitors (ICIs), either alone or combined with chemotherapy, represent the cornerstone of the treatment of advanced non-small cell lung cancer (NSCLC) without targetable gene alterations. Programmed death ligand-1 expression currently represents the only available biomarker to predict response to ICI, although its reliability is debated. However, most patients still do not derive benefit from immunotherapy, making the identification of further predictive biomarkers extremely needed. Serine/threonine kinase 11 (STK11)/liver kinase B1 (LKB1) and Kelch-like ECH-associated protein 1 (KEAP1) mutations occur in 25-30% and 11-27% of advanced NSCLC, respectively. Several studies associated their presence with poor outcomes in patients treated with ICI. However, more recent evidence showed poor outcomes among NSCLC with STK11/LKB1 and/or KEAP1 mutations regardless of the treatment received. We reviewed the literature to provide a comprehensive, timely and structured overview of the role of STK11/LKB1 and KEAP1 mutations in NSCLC. Although conflicting outcomes have been reported by studies evaluating their impact in KRAS wild-type patients or regardless of KRAS mutation, the correlation between STK11/LKB1 and KEAP1 mutations and poor outcomes with ICI appears to be consistent in presence of concurrent KRAS mutations. The main limitations of most studies are represented by the inclusion of other gene mutations (e.g. TP53) together with STK11 and KEAP1 mutations as a group and by the lack of comparison arms including patients who received other treatments (e.g. chemotherapy). Studies evaluating the impact of STK11 and KEAP1 mutations on the outcomes with ICI and other therapies showed a similar effect regardless of the treatment received, suggesting a prognostic, rather than predictive, value. (C) 2021 Elsevier Ltd. All rights reserved. 收起