Background: In the NALA phase III trial, irreversible pan-ErbB tyrosine kinase inhibitor (TKI) neratinib plus capecitabine demonstrated significant improvement in the progression-free survival (PFS), but no significant benefit in overall survival (OS) compared to lapatinib plus capecitabine. However, another TKI pyrotinib plus capecitabine showed significant benefits in PFS and a trend of OS benefits in the PHOEBE phase III study. But in general, current data on survival of irreversible TKIs in patients with HER2-positive metastatic breast cancer (MBC) was limited. Thus, we performed a pooled analysis of individual patient data from the phase I to III trials in HER2-positive MBC patients receiving pyrotinib or pyrotinib combined with capecitabine, to provide a cumulative assessment on long-term outcomes of irreversible TKI. Methods: Individual patient data was collected and analyzed from the phase I trial for pyrotinib and pyrotinib plus capecitabine, the Pivotal phase II trial and the PHOEBE phase III trials that enrolled patients in China National Cancer Center. Next-generation sequencing was performed on circulating tumor DNA for predictive biomarkers in the phase I trial. Results: Between August 2013 and October 2018, a total of 120 patients were assigned to received pyrotinib (n = 38), pyrotinib plus capecitabine (n = 53) and lapatinib plus capecitabine (n = 29) across the above four trials. The median follow-up duration for OS was 73.6 months (95% CI:69.9-77.3 months). The estimated median PFS was 8.2 months (95% CI:5.6-10.9 months) in the pyrotinib monotherapy group and 22.0 months (95% CI:13.2-30.8 months) in the pyrotinib plus capecitabine group (P= 0.002), while the median OS was 27.1 months (95% CI: 21.6-32.5 months) in the pyrotinib monotherapy group and 44.5 months (95% CI: 30.8-58.1 months) in the pyrotinib plus capecitabine group (P= 0.065). In this pooled analysis, pyrotinib 400mg in combination with capecitabine, recommended for phase II and III trials, had significantly longer PFS (22.0 vs 6.9 months, P＜0.001) and OS (59.9 vs 31.2 months, P= 0.035) than lapatinib plus capecitabine. Analysis of all genetic alterations in baseline blood samples suggested that the patients harbored concomitant mutations in HER2-related signaling network (including HER2 bypass signaling pathway, PI3K/Akt/mTOR pathway and TP53) were observed with significantly poorer PFS and OS compared to none or one genetic alteration (median PFS, 7.3 vs. 26.1 months, P= 0.003; median OS, 25.1 vs. 48.0 months, P= 0.013). Conclusions: This pooled analysis based on phase I to III trials revealed promising PFS and OS was achieved in pyrotinib and pyrotinib plus capecitabine. Concomitant mutations in HER2-related signaling network may be a potential efficacy and prognosis biomarker for pyrotinib in HER2-positive MBC. Clinical trial information: NCT01937689,NCT02361112,NCT02422199,NCT03080805.