Abstract
performance status, T and N stage, number of metastatic lesions, and EGFR mutation type were included in univariate analysis (UVA). Candidate factors on UVAwere incorporated into a multivariate analysis (MVA) to identify factors independently associated with BMFS. Results: Of 112 patients, 33 (29%) were male, median age at diagnosis was 60 years, and 35 (31%) had a history of smoking. Seventy-three (65%) and 34 (30%) carried exon 19 deletions and exon 21 (L858R) mutations, respectively; 5 patients had other mutations. In total, 38 patients developed BM, for a crude incidence of 33.9%. Median survival was 25.2 months, and among patients with BM, the median time to BM was 28.9 months. The cumulative incidence of BM at 2 and 3 years was 14.7% and 30.3%, respectively. On UVA, the only factor significantly associated with BM was L858R (exon 21) mutation (hazard ratio [HR] 1.93 vs exon 19 deletion, PZ.035). Older age was nonsignificantly associated with less BM (PZ.11). On MVA, L858R retained borderline significance (HR: 1.82, PZ.054). The cumulative incidence of BM at 2 years was 23.2% in the exon 21 group vs 10.4% in the exon 19 group. Conclusion: Brain metastasis is a significant pattern of failure in EGFRmutant metastatic NSCLC patients maintained on erlotinib therapy, occurring in approximately a third of such patients. Contrary to previous reports, we found that patients with L858R mutations may be at particularly greater risk of BM compared to patients with exon 19 mutations. These results suggest a potential role for PCI in this select group of patients as a means of improving CNS-progression free survival. Author Disclosure: S. Patel: None. A. Rimner: E. Research Grant; Varian Medical Systems. G. Consultant; General Electric. A. Foster: None. Z. Zhang: None. K.M. Woo: None. H.A. Yu: None. G.J. Riely: None. A.J. Wu: G. Consultant; Pfizer, Inc.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have