Systemic therapy and overall survival (OS) in patients (pts) with brain metastases from HER2-positive (HER2+) metastatic breast cancer (MBC).

Abstract

597 Background: Brain metastases (BM) occur frequently in Her2+ MBC. We have previously reported that trastuzumab (T) after diagnosis of BM may improve OS due to prolonged systemic disease control. Based upon those results we investigated whether lapatinib (L) may add additional survival benefit in such patients. Methods: 80 pts treated at the Medical University of Vienna for BM from Her2+ MBC were identified from a BC database. Karnofsky Performance Score (KPS) of at least 70 was required, as low KPS is a known negative predictor of OS. 36 pts were treated before 2002, when routine application of T after diagnosis of BM was started and served as control. The reminders received either T alone or T plus L (either concomitantly or sequentially) with or without chemotherapy after local therapy for BM. Her2 was determined by immunohistochemistry and reanalyzed by FISH if a score of 2+ was gained. OS from diagnosis of BM was defined as primary study endpoint and estimated using the Kaplan-Meier method. A Cox proportional hazards model was applied to correct for other factors associated with OS. Results: Median OS in pts receiving T after diagnosis of BM was 13 m (95%CI 8.85-17.15); corresponding numbers were 9 m (0-20.69) in pts treated with chemotherapy, and 3 m (2.37-3.63) with local therapy alone, respectively. Median OS was not reached in pts receiving L. Addition of L significantly prolonged OS over T alone (p=0.002). After correction for hormone receptor status, visceral metastases, early development of BM (<12 m), KPS, and number of BM (1-3 vs. >3), additional systemic therapy with L (HR 0.29; 95% CI 0.1-0.83; p=0.02) remained an independent predictor of better outcome; pts with >3 BM had significantly shorter OS (HR 6.56; 95%CI 2.44-17.67). Conclusions: Introduction of T improved OS in pts with Her2+ MBC. As our results suggest, addition of L may further improve OS in pts with BM. L and 1-3 BM were significantly associated with longer OS. KPS, on the other hand, had no significant influence, as the study population had a minimum KPS of 70 in order to avoid an inclusion bias. Therefore, pts with BM and higher KPS should receive all available systemic therapy options.