br Gene status EGFR ALK alteration is of significance
Gene status (EGFR/ALK alteration) is of significance for lung cancer patients with LM in the present study (HR: 0.5(0.4–0.8), p = 0.002). Consistent with other studies, patient with EGFR/ALK alteration, which were routinely tested nowadays, had a better OS than those with wild type [13,20]. Diverse gene status generally means to distinct targeted
Grouped Survival of Training Set and Validation Set in Lung Cancer Patients with LM.
Novel molGPA score Median Survival of Training set (m) Patient, No. (%) Median Survival of Validation set (m) Patient, No. (%)
group 1: high risk group group 2: immediate risk group group 3: low risk group.
Fig. 1. Kaplan-Meier curves for overall survival probability prediction of the training set: observed versus predicted overall survival. LM, leptomeningeal metastases. group 1: high risk group (molGPA score 0) group 2: immediate risk group (molGPA score 0.5–1.0) group 3: low risk group (molGPA score 1.5–2.0).
Fig. 2. Kaplan-Meier curves for overall survival probability prediction of the validation set: observed versus Predicted overall survival. LM, leptomeningeal metastases. group 1: high risk group (molGPA score 0) group 2: immediate risk group (molGPA score 0.5–1.0) group 3: low risk group (molGPA score 1.5–2.0).
therapies and clinical outcomes . There were few randomized trials conducted in lung cancer patients with LM who have actionable mu-tations. However, many retrospective studies revealed that EGFR TKI or ALK TKI had a promising anti-tumor activity for these patients [6–810,11,17,22]. [23,24] Our previous study showed that LM patients who were treated with EGFR TKI had a longer OS compared to those without TKI therapies (10.0 months vs. 3.3 months, P < 0.001), and this LOXO-101 findings were consistent with other similar studies. [6,10–12]. With a CSF cytological conversion rate of 64.3%, erlotinib (a first-generation EGFR TKI) showed high penetration rate to blood-brain barrier and promising anti-tumor effect against LM of lung cancer. Besides, the second and third generation EGFR TKIs such as afatinib, osimertinib and AZD3759 also demonstrated an efficacious effect [7,25,26]. When it comes to ALK TKI, studies on the treatment of ALK-positive lung cancer patients with LM remained scarce. With an excellent CNS
penetration, alectinib (a second-generation ALK/RET inhibitor) pro-duced radiological and neurological responses both in the conventional and high dose. [24,27] Overall, EGFR/ALK alteration plays an im-portant role in the LM of lung cancer.  Lacking of molecular as-sociated parameters, two classical models mentioned above may not be appropriate for lung cancer patients with LM, especially in the patients with actionable mutations.
In our model, the acceptable performance status (KPS score ≥60) was the strongest prognostic factor of better OS for lung cancer patients with LM (HR: 0.17(0.1-0.3), p < 0.001). Patients with KPS score 80–100 ha d a better OS than those with KPS 60–70 (Supplementary Data 4). Performance status (PS) was well recognized as the main prognostic factor in lung cancer patients with LM. Several retrospective studies showed that patients with a good Eastern Cooperative Oncology Group PS score (0–1) at the diagnosis of LM had a longer survival than those with a PS score of 2 or more. [6,8,14,16,17,29] KPS score ≥60 was also involved in the NCCN model, to classify LM patients into good risk and poor risk. However, NCCN model did not distinguish the dif-ferences between KPS score 60–70 and 80–100, which might weaken the discrimination capacity of the prognostic model.
Absence of ECM is a well-known predictor for better OS in brain metastases (BM) of lung cancer patients. [15,30,31] However, it was rarely mentioned in the prognostic studies on lung cancer patients with LM. In the present study, ECM showed a prognostic significance (HR: 0.6(0.4-0.8), p = 0.005). ECM was given a maximum of 0.5 based on its HR and statistical significance. The similar approach was used in the DS-GPA model and Lung molGPA model to predict survival in patients with BM [15,31].