The results of the test sets are de-identified, confidential, and securely stored in the BREAST databank for research and program evaluation purposes. A consent form is available to readers prior to doing a test set.
De-identified data are only made available to researchers through an application process with the BREAST Access and Management Committee (BAMC). Projects must have ethical approval and a completed access request form to be considered. The results arising from BREAST data will be non-re-identifiable and may be published in peer reviewed scientific journals.
The performance data of BreastScreen readers may be requested by the reader’s BreastScreen service leaders to monitor training activities. Your information will never be provided to other BreastScreen services or other states. BREAST provides confidential, anonymised activity reports to our funders on an annual basis, but these will never have identifying information.
BREAST treats medical images, reports and case information with utmost confidentiality and sensitivity. All images and reports that are used in BREAST’s systems undergo a strict anonymisation protocol. The de-identified images used in the BREAST test sets are collected from the BreastScreen digital libraries in Australia and New Zealand with ethical approval from the University of Sydney Human Research Ethical Committee.
All the cancer cases in our test set have gone through the BreastScreen program and have been verified by surgical pathology. All the cancer-free cases in our test sets have been determined by consensus reads by at least 2 senior radiologists in two rounds of normal screening.