A Prediction Model For Brain Metastasis In Non-small Cell Lung Cancer

Purpose:Lung cancer is the leading cause of tumor-related death worldwide and most of them are non-small cell lung cancer (NSCLC). Brain metastases are common in patients with non-small cell lung cancer and they are one of the most important causes of tumor relapse. However, nowthere is not any concise and accurate predictive method to judge which NSCLC patient may develop brain metastasis.In this study weinvestigatedthe association between EGFR and KRAS mutation and the risk of developing brain metastasis in a cohort of Chinese NSCLC patientsand made a risk prediction model fordevelopingbrain metastasis in NSCLCwiththe use of both genetic mutations and clinical covariates.Methods:Altogether191Chinese patients with NSCLCwho have accepted majority of their treatment at Peking Union Medical College Hospital were enrolled in this retrospectivestudy. EGFR and KRAS mutation status and Clinical information including gender, age, cell type, primary location of tumor, smoking history, tumor stage, lymph node metastasis at the time of diagnosis of NSCLCwere obtained and analyzed. Univariate and multivariate Cox regression model was used successivelyto investigatehow EGFR and KRAS mutations and clinical risk factors affected the risk ofbrain metastasis in NSCLC patients. A risk prediction model using genetic mutations and clinical risk factors with the parameter obtained in multivariate Cox regression analysis was established.Results:A total of26(13.6%) patients were diagnosed with brain metastases during the follow-up period. EGFR mutation (P=0.030), younger age (≤60years)(P=0.043) and lymph node metastasis (P=0.020) were found tosignif icantly increase the risk of developing brain metastasis among NSCLC patients. The prediction modelor risk prediction score for brain metastasis was made and high, intermediate, and low risk group were ranked. The3-year estimating of having brain metastasis for high, intermediate and low risk group was30.1%,20.5%and1.5%, respectively, stratifiedby EGFR mutation, age and lymph node metastasis(P<0.001).Conclusion:EGFR mutation, younger age≤60years and lymph node metastasis are significantlyassociated with the increasingrisk of developing brain metastasis in NSCLC patients. And the prediction model based on this discover is useful. In the future, NSCLC patients with positive EGFR mutation at high risk for brain metastasis should be considered for close surveillance. And trials of prophylactic therapy using targeted agents or prophylactic cranial irradiation (PCI) therapy should be developed.