This poster is published under an
open license. Please read the
disclaimer for further details.
Keywords:
Breast, Oncology, MR, MR-Diffusion/Perfusion, Computer Applications-Detection, diagnosis, Treatment effects, Cancer
Authors:
N. Michoux1, L. Bollondi2, A. Depeursinge3, A. Geissbuhler2, L. Fellah1, H. Müller3, I. Leconte1; 1Brussels/BE, 2Geneva/CH, 3Sierre/CH
DOI:
10.1594/ecr2015/B-1088
Results
Biological and imaging parameters
Neither the mass enhancement nor the non-mass enhancement were statistically different between NR and PR+CR.
NR were significantly more represented in Luminal-A subtype compared to PR+CR.
NR were significantly less represented in Ki67>14% and HR-/HER2+ compared to PR+CR (non-significant trend).
No statistical difference on histological grade between NR and PR+CR was observed.
Multi-parametric prediction
Computation parameters for texture analysis were: distance of one pixel between two neighboring pixels,
average of the angular relationships on the thirteen main directions,
four bits of grey levels.
Using SVM as classifier,
a predictive model relying on 3 Riesz parameters was found to perform with a predictive accuracy of 81%: Se = 47% (9/19 NR) and Sp = 94% (48/51 PR+CR).
Using the logistic regression as classifier,
a better model for identifying NR patients based on 5 textons (1 RLM + 4 Riesz) was found to perform with a predictive accuracy of 76%: Se = 89% (17/19 NR) and Sp = 71% (36/51 PR+CR).