Preliminary development of anatomically realistic breast tumor models for microwave imaging

In this work, we present a preliminary study of three classifiers - Linear Discriminant Analysis (LDA), Support Vector Machines (SVMs) and
$k$
-Nearest Neighbors (
$k$
NN) - to differentiate between malignant and benign tumors extracted from Magnetic Resonance (MR) images, based on their morphological features. The dataset in this study comprises 24 tumors: 12 malignant and 12 benign. Twelve morphological features were initially considered for tumor classification. The Mann-Whitney test was employed for feature selection, and the performance of the classifiers was evaluated with accuracy, sensitivity, specificity, F1-score and Matthew's Correlation Coefficient (MCC) metrics. kNN (with
$k=6$
and Chebyshev distance) outperformed the other classifiers with an accuracy, sensitivity, and specificity of 87.5%, 83.3% and 91.7%, respectively.