Molecular characteristics and prognostic factors of leptomeningeal metastasis in non-small cell lung cancer

Abstract

BackgroundNon-small cell lung cancer with leptomeningeal metastasis (NSCLC-LM) is emerging as a new management challenge for oncologists and is associated with poor prognosis. This study aimed to investigate the molecular characteristics and prognostic factors of NSCLC-LM. MethodsThis retrospective study included 97 patients with NSCLC-LM between January 2015 and October 2021. Progression-free survival (PFS) and overall survival (OS) were evaluated. Gene mutations were detected by next-generation sequencing (NGS). ResultsThe median PFS and OS were 8.4 (95 % confidence interval [CI]: 4.839–11.901) and 14.0 (95 % CI: 9.254–18.746) months, respectively. Sixty-seven patients harboured epidermal growth factor receptor-mutated (EGFRm): L858R (34), 19del (29), T790M (13), and G719C with L861Q (1). Other mutations included ALK (5), ROS1 (3), KRAS (1), TP53 (14), MET amplification (6). The detection rate and types of circulating tumour DNA (ctDNA) in the cerebrospinal fluid (CSF) samples were higher than the paired plasma samples. Patients with EGFR mutations had a longer median OS than those without mutations (19.0 vs. 13.0 months, P = 0.015). Patients with gene mutations had shorter median OS than those without mutations, such as ALK (11.8 vs. 19.9 months, P = 0.014), ROS1 (12.7 vs. 19.8 months, P = 0.014), KRAS (4.0 vs. 19.0 months, P = 0.005), TP53 (15.0 vs. 19.0 months, P = 0.014), and MET amplification (6.0 vs. 19.0 months, P = 0.003). Multivariate analysis indicated that MET amplification was an independent predictor of poor survival. Along with Eastern Cooperative Oncology Group Performance Status (ECOG PS) ≥ 3, LM accompanied with brain parenchymal metastasis (BPM), extracranial disease, and seizures were independent predictors of poor survival, whereas intrathecal chemotherapy, and third-generation EGFR-TKIs were independent predictors of favorable survival. ConclusionsCSF ctDNA detected using NGS had a high sensitivity for NSCLC-LM, showing high potential in detecting driver and drug-resistant gene mutations. Genomic profiles, combined with clinically relevant prognostic factors, will guide individualised treatments and improve the outcomes of NSCLC-LM patients.