MDNN: Predictive Modeling of Breast Cancer Subtypes Using Modified Deep Neural Networks
Kiran Chand Ravi, Sanjay Akaram Jadhav, R. Krishnamoorthy, K. Rajkumar, B. Swapna, Smita Sharma · 2023
Breast cancer remains a complex and heterogeneous disease, encompassing various molecular subtypes with distinct clinical implications. Accurate subtype classification is pivotal for tailoring treatment strategies and improving patient outcomes. In this research, we introduce a novel approach for predictive modeling of breast cancer subtypes, leveraging a Modified Deep Neural Network (MDNN) that combines Convolutional Neural Networks (CNNs) and GoogleNet architecture. Our study draws upon a comprehensive dataset that includes gene expression profiles and histopathological images from a large cohort of breast cancer patients. The MDNN architecture is meticulously crafted to harness the strengths of both CNNs and GoogleNet, allowing for the extraction of intricate features from histopathological images while effectively integrating gene expression data. The CNN component of MDNN excels at automatically capturing fine-grained patterns within histopathological images, enabling the model to discern subtle tissue nuances characteristic of different breast cancer subtypes. Simultaneously, the GoogleNet architecture enhances the model's ability to process and analyze high-dimensional gene expression data, facilitating a more holistic understanding of the underlying molecular characteristics. We employ rigorous data preprocessing, cross-validation, and hyperparameter tuning to train and optimize our MDNN model and obtained an accuracy of 98%. Comprehensive evaluation on an independent validation dataset showcases the superior performance of our approach in comparison to traditional methods.