Taxanes are the current first-line treatment for advanced cutaneous angiosarcoma (CAS) for patients who are considered difficult to treat with doxorubicin owing to advanced age or comorbidity. However, no effective second-line therapy for such patients has been established. We designed a single-arm prospective observational study of eribulin mesylate (ERB) administered at a dose of 1·4 mg m-2 on days 1 and 8 in a 21-day cycle. Patients with advanced CAS who were previously treated with a taxane and were scheduled to begin ERB treatment were enrolled. The primary endpoint was overall survival (OS) and the secondary endpoints were response rate (RR), progression-free survival (PFS) and toxicity assessment. We enrolled a total of 25 patients. The median OS and PFS were 8·6 months and 3·0 months, respectively. The best overall RR was 20% (five of 25). In total, 16 grade 3/4 severe adverse events (SAEs) occurred; however, all patients recovered. Patients who achieved partial response or stable disease as best response had longer OS than those with progressive disease (median OS not reached and 3·3 months, respectively; P < 0·001). Patients who did not experience SAEs showed longer OS than those who did (median OS 18·8 months and 7·5 months, respectively; P < 0·05). Patients with distant metastasis had shorter median OS than those with locoregional disease, but without statistically significant difference. ERB showed a promising RR and is a potential candidate for second-line treatment for patients with CAS, after treatment with taxanes. However, owing to the occurrence of SAEs in over half of the participants, caution should be exercised regarding ERB use in elderly patients. What is already known about this topic? Taxanes are the current first-line treatment for patients with advanced cutaneous angiosarcoma (CAS) who are considered difficult to treat with doxorubicin owing to advanced age or comorbidity. No effective therapy for taxane-resistant CAS has been established thus far. Eribulin suppresses microtubule polymerization and elicits an antitumour effect similar to that of taxanes. What does this study add? In our single-arm prospective observational study to evaluate the efficacy of eribulin for treating patients with advanced CAS who previously received taxanes, the median overall survival and progression-free survival were 8·6 and 3·0 months, respectively. Response rates at weeks 7, 13 and 25 were 20%, 17% and 14%, respectively. Although 16 grade 3/4 severe adverse events occurred, all patients recovered. Eribulin showed a promising response rate and is a potential candidate for second-line treatment in CAS after taxane treatment. Linked Comment: Smrke and Benson. Br J Dermatol 2020; 183:797-798. © 2020 British Association of Dermatologists.