Textural Analysis of DCE-MRI of the Breast as a Predictor of Response

Peter Gibbs¹, Arfan Ahmed¹, Martin Pickles¹, Lindsay Turnbull¹

Textural analysis, based on the spatial grey-level dependence matrix method, has been performed on 100 breast cancer patients prior to treatment with neoadjuvant chemotherapy. Texture parameters were calculated pre-contrast administration and 1, 2, 3, 4 and 5 minutes post-contrast injection. Significant differences in f2 (contrast) and f10 (difference variance) were noted between partial responders and non-responders. Significant differences were also noted in f6 (sum average) and f15 (cluster shade) between node positive and node negative patients. Textural analysis of dynamic contrast enhanced data may have a role in predicting treatment response.