A Machine Learning Validation for Identifying the difference between Benign Cysts and Malignant Tumor in Breast Cancer

Abstract Breast cancer is a leading health concern for women globally. Hormonal fluctuations can lead to benign breast abnormalities, such as cysts, such as cysts, which may have potential to become cancerous. Researchers have been proposed various thermal modelling techniques to detect breast cancer by differentiating between healthy and cancerous breast tissues. There is a gap literature regarding the use of thermal models to differentiate between benign and malignant breast disorders in women. This research focuses on the thermal changes caused by benign and malignant breast tumor in fatty tissue of woman’s breast. Experiment was conducted in School of Advanced Sciences and languages, VIT Bhopal University, India. That was approved by King Khalid University, and University of Arizona. A 2D steady-state model was developed using Penne’s bioheat equation, incorporating ket parameters such as blood flow rate, thermal conductivity and metabolic heat generation. The model incorporates adiabatic boundary conditions were defined for different environmental conditions. Solutions were computed using finite element analysis. Simulations yielded results for various spherical cysts and different tissue depths within a hemispherical breast model. Furthermore, this study compared thermal profiles associated with cysts and malignant tumors breast tissue. This study reveals distinct thermal differences between cyst and malignant tumour in breast tissue, potentially aiding in accurate diagnosis and reducing false-positive. This study identified several factors influencing cancer classification and prediction accuracy. The data used here is publicly available on Mammographic Image Analysis Society (MIAS) database [35]. Cysts and malignant tumors exhibit opposite thermal effects in breast tissue. The curve’s slope at the cyst-normal tissue interface is opposite to that at the tumor-normal tissue interface. Thermography can leverage the distinct thermal signatures of cysts and tumors to detect and characterize these abnormalities in breast tissue. Thermography can accurately differentiate cysts and cancerous tumors, reducing false positives. Notably, this technique can detect small lump (2mm diameter/1mm radius) at depths of up to 2.5 cm. This study finds SVM outperforms RF and ANN in breast cancer classification, leveraging computationally extracted features (viz., avg. accuracies of SVM, ANN and RF are 96.66%, 93.88% and 90.55%, respectively). In this study, we classified breast cancer data using three machine learning techniques: Support Vector Machine (SVM), Artificial Neural Network (ANN) and Random Forest (RF). We optimized model performance by training with selected features based on the analysis.