1305P - Factors predicting brain metastases in patients with non-small cell lung cancer

Date 29 September 2012
Event ESMO Congress 2012
Session Poster presentation I
Topics Non-Small-Cell Lung Cancer, Metastatic
Presenter Shih-Hsin Hsiao
Authors S. Hsiao1, C. Chung1, H..E. Liu2
  • 1Division Of Pulmonary Medicine, Department Of Internal Medicine, Taipei Medical University Hospital, Taipei, 110 - Taipei/TW
  • 2Division Of Hematolgy And Oncology, Department Of Internal Medicine, Department of Medicine, Wanfang Hospital, Taipei Medical University, 111 - Taipei/TW



Brain metastases (BM), a common complication of non-small cell lung cancer (NSCLC), usually lead to a poor prognosis. Recent advances in BM therapy modesly prolong the survival after BM diagnosis in a subset of patients. Selection of treatment modalities for BM is based largely on the number of BM, BM-related symptoms and patient's functional performance status. Therfore, early dection of BM in high-risk patients is crucial. In this study, we sough to elucidate the factors predicting BM.

Methods and patients

Medical records of patients with stage 1-4 NSCLC were retrospectively reviewed for the period between January 2006 and December 2011 under the approval of the joint institutional review board. Clinical demographic data, histology, stage of disease, presence of BM, survial were collected and analyzed. A multivariate logistic regression model was used to identiy the predictors of BM.


Among 596 NSCLC patients with a mean follow-up time of 12.5± 12.5 months and a mortality rate of 62% at the last follow up, 187 (31%) experienced BM during their disease course. The accumulative incidence of BM was higher in patients with adenocarcinoma (ADC) than those with squamous cell carcinoma (SCC) (36% vs 13%, p < 0.001). On multivatiate analysis, female, age < 60 years, ADC and stage 3b/4 were significantly associated with BM (OR = 1.67, 95% CI = 1.06-2.63, p = 0.025, and OR = 1.89, 95% CI = 1.28-2.78, p = 0.001, and OR = 2.67, 95% CI = 1.35-5.26, p = 0.017, and OR = 3.94, 95% CI = 2.15-7.13, p < 0.001, respectively). The incidence of BM in stage 3b/4 NSCLC patients was 36% and varied from 14% to 59% in patients with or without identified risk facotrs. Specifically, ADC patients with age less than 60 years were more likely to experience BM than the elder with SCC (OR = 5.46, 95% CI = 2.79-10.71, p < 0.001). Additionally, prolonged survial after diagnosis of lung cancer heightened the risk of BM; its incidence was 42%, 54% and 64% in patients who survived longer than 3, 12 and 24 moths, respectively.


We find that gender, age and histological subtype are independent factors predicting BM in NSCLC patients and suggest identification of patients at high risks for BM might help detect BM earlier and facilitate the design of clinical trials aming at the prevention of BM.


All authors have declared no conflicts of interest.