Meeting Banner
Abstract #1164

Predicting breast cancer patient survival using textural features in CE MR images

Beathe Sitter1, Guro Fanneløb Giskeødegård1, Ioanna Chronaiou1, Jose Teruel2, Roja Hedayati3,4, Steinar Lundgren3,4, Else Marie Huuse Røneid5, Martin Pickles6, Peter Gibbs7, and Tone Frost Bathen1

1Department of Circulation and Medical Imaging, NTNU, Trondheim, Norway, 2Department of Radiation Oncology, NYU Langone Health, New York, NY, United States, 3Cancer clininc, St. Olavs University Hospital, Trondheim, Norway, 4Department of Clinical and Molecular Medicine, NTNU, Trondheim, Norway, 5Department of Radiology, St. Olavs University Hospital, Trondheim, Norway, 6Radiology department, Hull & East Yorkshire Hospitals NHS Trust, Hull, United Kingdom, 7Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY, United States

Treatment for women with locally advanced breast cancer (LABC) is determined with inadequate knowledge of the long-term outcome. We evaluated the prognostic value of textural features derived from pre-treatment CE-MRI in 55 LABC patients scheduled for neoadjuvant chemotherapy. Using overall survival at 7-years follow-up as endpoint, textural features derived from post-contrast pre-treatment images were significantly different. Using all textural features as input for multivariate analysis, we achieved a classification accuracy of 72% (p<0.001), which increased to 78% when including traditional prognostic factors (p<0.001). Textural features provide prognostic information, which can complement the stratification of patients to treatment.

This abstract and the presentation materials are available to members only; a login is required.

Join Here