Background: Despite the large number of patients with early breast cancer (EBC) that have been treated with capecitabine in randomized trials no individual patient data meta-analysis has yet been conducted. So far only two literature-based meta-analyses have been published including only five neoadjuvant studies and two adjuvant trials, respectively. Whereas the first did not report an improvement in response to neoadjuvant therapy, the latter found an improvement in disease-free survival (DFS) in the capecitabine arm.

Methods: In order to select trials the following criteria were used at Use of capecitabine in EBC as adjuvant or neoadjuvant therapy; randomized; N>100 patients; recruitment completed and outcomes available. This yielded the following trials: GeparTrio, GeparQuattro, GAIN, ICE, ICE II, FinXX, Alliance/CALGB 49907, Roche / US Oncology 01062, NSABP-B40, ABCSG-24, CREATE-X, GEICAM 2003/10, GEICAM 2003/11_CIBOMA 2004/01. Data were available for all these trials and 15,457 patients were included in this individual patient data based meta-analysis. The primary objective was to examine the effect of capecitabine on disease-free survival (DFS). Secondary objectives were to examine effects on distant DFS (DDFS), overall survival (OS), pathological complete response (for neoadjuvant studies) and whether there was an interaction of occurrence of capecitabine-specific toxicity (mucositis, diarrhea, hand-foot syndrome) and treatment effect. Bi- and multivariable frailty-based Cox proportional hazards models including log-normal distributed random effects of study were used to report hazard ratios with 95% CI between the groups of patients treated with or without capecitabine for DFS, DDFS and OS.

Results: Individual data from 15,457 patients was collected. Of these 7,980 received capecitabine during the course of their treatment and 7,477 patients were treated in the control arms. Median age at diagnosis was 54 years in both groups. Most patients were diagnosed with stage 2 tumors (55.9%) and the majority presented with nodal involvement (74.0%). Estrogen and progesterone receptor positivity was observed in 66.0% and 56.9%, respectively, and 15.1% of patients were diagnosed as HER2-positive. 2,816 patients (18.2%) received neoadjuvant treatment and 12,641 (81.8%) an adjuvant chemotherapy regimen. Cox regression analyses of all included patients revealed that the addition of capecitabine did not alter DFS significantly compared to treatment without capecitabine (HR 0.952; 95% CI 0.895-1.012; p-value 0.115). There was also no effect on DFS if studies were selected in which capecitabine was given instead of another drug (HR 1.035; 95% CI 0.945-1.134; p=0.455). However, capecitabine in addition to systemic treatment lead to a small but significant improvement of DFS (HR 0.888; 95% CI 0.817-0.965; p=0.005). Results concerning other endpoints will be presented at the meeting.

Conclusion: Capecitabine did not alter DFS in this meta-analysis of 15,457 patients with early breast cancer from 12 prospective randomized trials, but as addition to systemic treatment DFS was improved. Subgroup analyses are needed for final interpretation and will be presented at the meeting.

Citation Format: Marion van Mackelenbergh, Fenja Seither, Volker Möbus, Joyce O'Shaugnessy, Miguel Martin, Heikki Joenssuu, Michael Untch, Ulrike Nitz, Juan J Miralles, Masakazu Toi, Harry D Bear, Hymann Muss, Toralf Reimer, Valentina Nekljudova, Sibylle Loibl. Effects of capecitabine as part of neo-/adjuvant chemotherapy. A meta-analysis of individual patient data from 12 randomized trials including 15,457 patients [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr GS1-07.