Influence of the Timing of Leptomeningeal Metastasis on the Outcome of EGFR-Mutant Lung Adenocarcinoma Patients and Predictors of Detectable EGFR Mutations in Cerebrospinal Fluid

SIMPLE SUMMARY: Leptomeningeal metastasis (LM) is a devastating complication of lung cancer, with a generally poor outcome. We conduct the present study to evaluate the association between clinical presentations, brain images, tumor cell counts of the cerebrospinal fluid (CSF), and the epidermal growth factor receptor (EGFR) mutation detection rate in CSF among EGFR-mutant lung adenocarcinoma patients with LM and accessed the influence of the timing of LM occurrence on patient outcomes. Tumor cell numbers were semi-quantified according to tumor cells per high power field of CSF cytological slides. Radiological burden was assessed using a four-point scoring system, which evaluated LM-involved areas on brain magnetic resonance imaging. Our results suggest the association between the radiological severity score of LM, CSF tumor cell counts, and EGFR mutation detection rate in CSF. Furthermore, LM prior to first-line EGFR-tyrosine kinase inhibitor treatment was associated with an independently worse outcome. ABSTRACT: Background: We aim to evaluate the influence of the timing of leptomeningeal metastasis (LM) occurrence on the outcome of EGFR-mutant lung adenocarcinoma and to explore the predictors of detectable EGFR mutation in the cerebrospinal fluid (CSF). Methods: EGFR-mutant lung adenocarcinoma patients with cytologically confirmed LM were included for analysis. EGFR mutation in CSF was detected by MALDI-TOF MS plus PNA. Results: A total of 43 patients was analyzed. Of them, 8 (18.6%) were diagnosed with LM prior to first-line EGFR-TKI treatment (early onset), while 35 patients (81.4%) developed LM after first-line EGFR-TKI treatment (late onset). Multivariate analysis suggested that both late-onset LM (aHR 0.31 (95% CI 0.10–0.94), p = 0.038) and a history of third-generation EGFR-TKI treatment (aHR 0.24 (95% CI 0.09–0.67), p = 0.006) independently predicted a favorable outcome. EGFR mutation detection sensitivity in CSF was 81.4%. The radiological burden of LM significantly correlated with CSF tumor cell counts (p = 0.013) with higher CSF tumor cell counts predicting a higher detection sensitivity of EGFR mutation (p = 0.042). Conclusions: Early onset LM was an independently poor prognostic factor. A higher radiological severity score of LM could predict higher tumor cell counts in CSF, which in turn were associated with a higher detection rate of EGFR mutation.