The QIN-Breast Dataset consists of longitudinal PET/CT images and directional MR images, which are used to study new methods of adjuvant treatment of breast cancer. It contains 100,835 images of 68 patients. The dataset covers images in three time periods: before the start of treatment T1; after one cycle of treatment T2; after the second cycle of treatment or after all treatments are completed T3.

The value of this dataset lies in clinical imaging data, which can be used to develop and evaluate quantitative imaging methods for early evaluation of breast cancer treatment. The data were provided by Vanderbilt University – PI Dr. Thomas E. Yankeelov, and the equipment was a GE Discovery STE scanner. The acquisition data of the scanning CT are as follows: for patients weighing 70kg, the tube current is 80mAs, the tube voltage is 120kVp, the spacing is 1.675/1, and the FDG administration activity is about 370MBq. The emission data of each bed position was collected in 3D mode after 1 hour, with a time interval of 2 minutes. Initially, only the breast prone position was collected, and later the supine position data from the skull to the mid-femur were collected.

This dataset was released by the Cancer Imaging Archive (TCIA) in 2016. The related paper is “Data From QIN-Breast. The Cancer Imaging Archive”.