Fig 1.
An example of a breast DCE-MRI slice.
(a) and its CAD-segmented breast regions depicted on the image slice before contrast agent injection (S-0) in which the non-breast regions behind the chest wall are removed.
Table 1.
Summary of extracted DCE-MR image features.
Table 2.
Characteristics of the molecular subtypes of patients.
Fig 2.
Representative imaging features differentiate tumors with different molecular subtypes.
A) Postcontrast MRI of luminal A and luminal B breast cancers. B) Postcontrast MRI of luminal A and luminal B breast cancers and corresponding kurtosis density using a kernel smoothing function.
Table 3.
List of 24 features that were selected in leave-one-out training and testing cycles to test 60 cases.
Fig 3.
ROC curve in the first cohort for classifying the four molecular subtypes of breast cancer.
The classifiers based on dynamic features, morphologic features, first-order statistical features and clinical information are shown. Features are combined to classify between (a) luminal A and non-luminal A tumors; (b) luminal B and non-luminal B tumors; (c) HER2-positive and non-HER2-positive tumors; and (d) basal-like and non-basal-like tumors.
Table 4.
List of 15 image features that were selected in leave-one-out training and testing cycles to test 36 cases in the validation dataset.