Effects of taxanes alone or in combination with anthracyclines on tumor response, progression-free survival and overall survival in first-line chemotherapy of patients with metastatic breast cancer: an analysis of 4,256 patients randomized in 12 trials

Abstract

Background: Taxanes (T, paclitaxel or docetaxel), in view of their partial lack of cross resistance with anthracyclines (A, doxorubicin or epirubicin), generated great enthusiasm in the 1990’s for the treatment of breast cancer. Taxanes were sequenced or combined with anthracyclines and compared with standard regimens for first-line treatment of advanced disease. Randomized trials, however, have shown inconsistent results, particularly in terms of survival, which was rarely improved. This prompted the current meta-analysis of individual data from all relevant trials, in order to reliably detect any advantages of taxanes in terms of tumor response, progression-free survival (PFS) and overall survival (OS) in first-line treatment of patients with metastaic breast cancer. Methods: Between December 2002 and May 2005, individual patient data were collected on 9 randomized combination trials comparing A + T (+ cyclophosphamide in 1 trial) with A + cyclophosphamide (+ 5-fluorouracil in 4 trials), and on 3 single agent trials comparing T (paclitaxel in 2 trials) with A. Combination trials included 3,337 patients, single agent trials 919 patients. Tumor response was assessed using WHO criteria. Progressionfree survival was calculated from date of randomization to progression or death from any cause. Results: Median follow-up of alive patients was 41 months, median OS 20 months, median PFS 7 months. In single agent trials, overall response rates were 33% (4% complete) in the T arms and 38% (6% complete) in the A arms (P=0.08). The hazard ratios for T compared with A were 1.19 (CI 1.04 – 1.36, P=0.01) for PFS and 1.01 (CI 0.88 – 1.16, P=0.90) for OS. In combination trials, response rates were 56% (10% complete) in T-based combinations and 45% (6% complete) in control arms (P<0.001). The hazard ratios for T-based combinations compared with control arms were 0.93 (CI 0.87 – 1.00, P=0.06) for PFS and 0.95 (CI 0.88 – 1.03, P=0.23) for OS. Conclusions: Single agent A was significantly better than single agent T in terms of PFS, marginally better in terms of response but not different in terms of OS. T-based combinations were significantly better than A-based combinations in terms of response rates, marginally better in terms of PFS but not different in terms of OS.