Skip to main content
Fig. 4 | BMC Cancer

Fig. 4

From: Texture analysis on MR images helps predicting non-response to NAC in breast cancer

Fig. 4

Pixel-level analysis of breast MRI texture in a CR patient with a mass enhancement. Are respectively displayed, a the axial subtracted image and the maps based on b contrast, c correlation, d difference variance, e energy, f entropy, g inverse differential moment (which is correlated with the homogeneity parameter), h sum average and i sum variance from the GLCM, with mean value estimated on a 3x3 neighbourhood around the pixel of interest then normalized on the 0–255 range. Individual texture parameters reveal different local and regional statistical properties of the grey level intensity between (and respectively within) breast lesions and normal parenchyma. Combination of all or parts of the texture parameters helps classifying patients according to their response to NAC

Back to article page