P2.16-09 Patterns of Brain Metastases in Stage I-III Non-Small Cell Lung Cancer: An NCDB Analysis

Abstract

We have submitted single institution data suggesting nodal size is most associated with the development of brain metastases at diagnosis of non-small cell lung cancer, and seek to investigate the applicability of NCCN recommendations for brain magnetic resonance imaging (MRI) for non-small cell lung cancer (NSCLC) to evaluate rates brain metastases in patients with Stage I-III non-small cell lung cancer (NSCLC) in a population based cohort. Results of patients diagnosed with NSCLC in the NCDB were evaluated between the years 2010 and 2013. Patients were included if they were AJCC 7th edition Stage I-III, or Stage IV patients on the basis of brain metastases alone. Exclusion criteria included those with bone, lung, or liver metastases at diagnosis. Due to heterogeneity in reporting in the NCDB, characterization of Stage IB, IIA, and IIB is difficult to discern. To account for this, we present rates of brain metastases for patients with T2N0, T1/2N1, and T3N0 disease within this cohort. Statistical evaluation included incidence of brain metastases per stage as well as logistic regression of baseline factors predictive of brain metastases (SPSS V21). A total of 256,340 patients were included in the analysis. Overall rates of brain metastases were 6.4% (16,333/256,340). Brain metastases were identified in 1.6% (1457/89,868) of Stage IA patients, 5.5% (4038/73,233) of patients with Stage IB-II disease (4.6% (2025/44,026), 8.2% (1162/14,118), and 5.6% (851/15,089) for stages T2N0, T1/2 N1, and T3N0, respectively), 11.0% (6945/63,127) of Stage IIIA patients, and 12.9% (3893/30,112) of Stage IIIB patients. On multivariate analysis, younger age was associated with increased risk of brain metastases (p<0.01). When compared to Stage IIIB, patients with T1/2 N1 disease and Stage IIIA disease were most likely to have brain metastases (p<0.01 for both and HR 0.823 and 0.959, respectively). Poorly differentiated tumors were most associated with brain metastases (p<0.01 for both; HR 0.652 vs 0.187). Additionally, women were more likely to have brain metastases than men (p<0.01, HR 1.065). Adenocarcinoma histology was most likely to have brain metastases in our population (p<0.01, HR 1.162). This series shows that young age, adenocarcinoma histology, female gender, higher clinical stage disease, and younger age were associated with the development of brain metastases. Future analysis will be focused on developing a RPA classification of this data to further prognosticate most important features associated with the development of brain metastases in newly diagnosed NSCLC.