Multidisciplinary management of patients with non–small cell lung cancer with leptomeningeal metastasis in the tyrosine kinase inhibitor era

OBJECTIVE Leptomeningeal metastasis (LM) is a challenging scenario in non–small cell lung cancer (NSCLC). Considering that outcomes of treatment modalities stratified by LM chronological patterns related to brain metastasis (BM) are lacking, the aim of this study was to evaluate outcomes and explore prognostic factors. METHODS The authors retrospectively collected data of patients with NSCLC undergoing Ommaya reservoir implantation, ventriculoperitoneal shunt implantation, or lumboperitoneal shunt implantation. Based on radiographic findings and time from diagnosis of NSCLC to LM, the authors divided them into subtypes of LM as follows: LM without BM; LM concurrent with BM; or LM after BM. The Kaplan-Meier method was applied to analyze overall survival (OS) and multivariate Cox regression for prognostic factors. RESULTS Sixty-one patients with LM were included, with a median OS of 8.1 (range 0.2–70.0) months. Forty-three (70.5%) patients had EGFR -mutant disease. Forty-two (68.9%) patients had 19-del or L858R mutation, and one (1.6%) patient had G719A mutation. Fifty-seven (93.4%) patients had hydrocephalus. Twenty-one (34.4%) patients received whole-brain radiotherapy before LM diagnosis, 3 (4.9%) patients underwent operation for BMs before LM diagnosis, and 42 (68.9%) patients received EGFR tyrosine kinase inhibitor (TKI) therapy before LM diagnosis. Eleven patients were treated with chemotherapy, 10 patients were treated with TKIs, and 32 patients were treated with chemotherapy combined with TKIs before LM diagnosis. Patients with LM after BM had lower Karnofsky Performance Status (KPS) scores (KPS score 50) than did those with LM without BM (KPS score 80) or LM concurrent with BM (KPS score 70; p = 0.003). More patients with LM after BM received intrathecal methotrexate than in the other subgroups (p < 0.001). The median OS was significantly shorter in the LM after BM than in the concurrent LM and BM and the LM without BM subgroups (5.4 vs 5.5 vs 11.6 months; p = 0.019). Cox regression revealed that a KPS score ≥ 70 (HR 0.51; p = 0.027) and shunt implantation (HR 0.41; p = 0.032) were favorable prognostic factors. CONCLUSIONS Patients with NSCLC who had LM without BM had better survival outcomes (11.6 months) compared with those who had LM after BM or concurrent LM and BM. Aggressive shunt implantation may be favored for LM.