The mechanism by which trastuzumab-emtansine (T-DM1) causes systemic toxicities apart from trastuzumab alone is currently unknown. We hypothesized that the systemic toxicities from T-DM1 may have been caused by the free and active maytansine released from the lysed HER2+ tumor cells, and if so, they may correlate with the response to treatment and eventually disease-free survival or patient outcome. In a retrospective, observational study, we evaluated 73 patients from three centers in the United States and Canada with advanced HER2+ breast cancer that received at least one dose of T-DM1. Toxicity grades were summed to create a corresponding toxicity sum score (TSS), and its association with clinical outcomes was analyzed. A higher TSS was significantly associated with longer progression-free survival with an HR = 0.66 [95% confidence interval [CI]: 0.47-0.92], P =.014, for each 1-point increase in the TSS score. Adjusted for baseline platelet count, aspartate transaminase and alanine transaminase, higher TSS remains significantly associated with longer progression-free survival with adjusted HR = 0.67 [95% CI: 0.47-0.93], P =.020. The analysis suggests that the systemic toxicities of T-DM1 were significantly correlated with its clinical efficacy. This is the first report to correlate the systemic toxicities of T-DM1 with clinical outcome. Further, this suggests that systemic toxicities of antibody-drug conjugates (ADCs) may serve as a predictive biomarker, particularly if noncleavable linkers are used. If confirmed in larger prospective studies, the present finding is significant because most ADCs do not have a biomarker predictive of clinical outcome other than the presence or absence of the antibody target.