Our previous study showed that LM were much more frequently
Our previous study showed that LM were much more frequently seen in EGFR- mutated patients.  And patients with EGFR mutations have a better OS than those with EGFR wild types.  Due to the existence of blood-brain barrier, many treatments in the past were often considered futile for central nervous system metastasis, especially me-ningeal metastasis . But, recent molecular therapies have shown high penetration to blood-brain barrier and promising antitumor effects in lung cancer patients with LM [6,7,10–1316–18]. Thus, in the context of targeted therapies, those two models without integration of mole-cular parameters might not be appropriate for estimating the prognosis of lung cancer patients with LM, especially in the patients with ac-tionable mutations.
We performed this study to establish a lung cancer specific graded prognostic assessment model with integration of molecular alterations (molGPA) to estimate survival in lung cancer patients with LM.
2. Materials and methods
We retrospectively screened 8921 consecutive lung cancer patients at Guangdong Lung Cancer Institute (GLCI) from January 2011 to March 2018. Among these patients, we identified 336 lung cancer pa-tients (3.8%) who were diagnosed as LM in our institute. The diagnostic criteria of LM were CSF positive for tumor cells, or CSF liquid biopsy positive for EGFR mutations or ALK rearrangements, or typical imaging of Benazepril MRI (linear or micro-nodular pial enhancement in gadolinium-enhanced brain MRI) [7,14,19]. Twenty patients who were lost to follow-up and fifteen patients who had no sufficient electronic medical records in our hospital were excluded from this study. As a result, we analyzed 301 patients who received any kind of anti-cancer treatment in our hospital and had sufficient clinical information (baseline demo-graphics, clinical presentations, and treatment course). They were randomly divided into training set (200 patients) and validation set (101 patients) after stratified by gender and age. This retrospective study was approved by the Hospital’s Research Ethic Committee, and complied with the regulations of clinical studies.
The electronic medical records of the enrolled patients were re-viewed, including their demographic data, laboratory and imaging ex-aminations, clinical presentations, and major treatment course for LM. Examinations included their histology types, gene profiles of EGFR mutation and ALK alteration, Thinprep cytologic test (TCT) and brain MRI. Treatments including Tyrosine Kinase Inhibitor (TKI), che-motherapy, radiotherapy such as WBRT and WBI were recorded in this study. Therapeutic regimens for most cases were carried out after multidisciplinary discussions in our institute. Continuous variables were categorized based on clinical reasoning and statistical methods. Age was grouped as > 63 years or ≤63 years. KPS status was grouped as a score of ≥60 or < 60.
2.3. Statistical analysis
OS was defined from the date of initial diagnosis of LM to the date of death or censored at the date of last follow-up (November 14, 2018). Variables were compared by the chi-square test and Fisher’s exact test. Survival was estimated with the Kaplan-Meier method and compared by the log-rank test between different types. We used multiple Cox Lung Cancer 131 (2019) 134–138
Characteristics of the Training and Validation Sets of the 301 Lung Cancer Patients with Leptomeningeal Metastases.
Variables Training set Validation Set P-value
Treatments before diagnosis
a Thinprep cytologic test. b EGFR/ALK TKI sensitive mutation.