AIM: To assess whether ultrasound features of breast cancer are associated with breast cancer specific survival (BCSS). MATERIALS AND METHODS: Within a single breast service, data was collected prospectively (April 2010–April 2012) from 319 consecutive women (mean age 63 years) with 335 ultrasound-visible invasive breast cancers. Ultrasound features were evaluated retrospectively from the recorded images according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon by a radiologist blinded to outcomes. Survival and cause of death were ascertained from local and national sources. Kaplan–Meier survival curves were generated, and statistical significance tested using the log-rank test. RESULTS: Mean follow-up in those alive was 80.9 months. Thirty breast cancer deaths and 45 non-breast cancer deaths occurred. Five-year BCSS in the presence of distal acoustic enhancement was 76% compared to 88%, 96%, and 100% for those with distal shadowing, no distal effect or combined effect respectively (p<0.0002). Patients with sonographic skin involvement had 73% 5-year BCSS compared to 92% for no skin involvement (p<0.0001). Focal oedema was associated with 56% 5-year BCSS compared to 89% for those without (p=0.0002). A significant association was demonstrated between ultrasound tumour size and BCSS (p<0.0001). At multivariate analysis, skin changes, distal enhancement, and focal oedema maintained prognostic significance. CONCLUSION: Distal enhancement, focal oedema, and skin involvement have strong associations with breast cancer death. These factors could be taken into account, along with lesion size and other commonly used preoperative prognostic features, when considering management of women with breast cancer.