Xiaoxing Gao, Jie Gao, Qiuyue Ye, Chen Wei, Qing Zhou, Xiaoyan Liu, Minjiang Chen, Jing Zhao, Hanping Wang, Xiaoyan Si, Wei Zhong, Yan Xu, Mengzhao Wang
The later era was associated with a longer interval to LM diagnosis but no statistically significant improvement in unadjusted post-LM OS. Adjusted associations require cautious interpretation because of residual confounding and post-baseline treatment bias. CSF profiling may provide clinically relevant information beyond primary tumor genotyping.
INTRODUCTION: Leptomeningeal metastasis (LM) is a devastating complication of lung cancer. We investigated the clinical and molecular evolution of LM in the modern targeted-therapy era and factors associated with post-LM survival.
METHODS: We retrospectively analyzed 201 patients with cytologically confirmed lung cancer LM diagnosed between November 2012 and May 2026. Patients were classified by LM diagnosis date (2012-2016 vs. 2017-2026). Intervals from lung cancer and Stage IV diagnosis to LM were compared using the Mann-Whitney U test. Paired primary tumor and cerebrospinal fluid (CSF) molecular profiles were assessed in 80 patients. Overall survival (OS) was evaluated using Kaplan-Meier and prespecified multivariable cox analyses.
RESULTS: Adenocarcinoma accounted for 93.0% of cases. The intervals from lung cancer and Stage IV diagnosis to LM were longer in the later cohort (23.7 vs. 13.8 months, p < 0.001; 17.4 vs. 10.4 months, p = 0.002). Unadjusted OS from LM diagnosis did not differ significantly between eras (9.8 vs. 11.3 months, p = 0.864). Tissue-CSF driver discordance occurred in 7/80 patients (8.75%). In multivariable analysis, later diagnostic era, smoking, and ECOG PS ≥ 2 were associated with higher mortality (adjusted HR, 1.826, p = 0.014; 1.847, p = 0.014; and 1.674, p = 0.007, respectively), whereas post-LM third-generation EGFR-TKI exposure was associated with lower mortality (adjusted HR, 0.385; p < 0.001).
CONCLUSION: The later era was associated with a longer interval to LM diagnosis but no statistically significant improvement in unadjusted post-LM OS. Adjusted associations require cautious interpretation because of residual confounding and post-baseline treatment bias. CSF profiling may provide clinically relevant information beyond primary tumor genotyping.