Integrating Imaging and Genomic Data with Composite AI to Enhance Breast Cancer Diagnosis and Early Detection

Abstract

Early detection of breast cancer will increase patient outcomes and survival, which will aid in improved patient care. However, integrating genetic data with medical imaging will enhance our understanding of the condition, resulting in a more accurate diagnosis and tailored therapy for the patient. Multiple artificial intelligence techniques coupled to create composite AI can enhance the interpretation of the aggregated data and pave the way for new, cutting-edge research directions in breast cancer. In this study, we present an integrated artificial intelligence system that improves breast cancer detection by combining data from genomics and imaging techniques. Our solution is a hybrid strategy that employs CNN models for image data and machine learning models (SVMs and Random Forests) for processing genomic data. We employ stacking, an additional ensemble learning technique, to merge the predictions of the CNN and genomic data models. As a result, our method produces a reliable and accurate diagnosis model by capturing unique imaging and genetic data. By capturing characteristics unique to picture data and genetic traits, this hybrid technique enables a more potent prediction model. Datasets that included genetic profiles with mammography pictures obtained from publicly accessible depositories such as TCGA and TCIA were used to evaluate the algorithms. Using the imaging data, the CNN model demonstrated an 88% accuracy rate in distinguishing between benign and malignant samples. An accuracy of 85% was obtained by analyzing the genomic data using SVM findings. With a classification accuracy of 92%, the composite model, which combined the output of the CNN and SVM models, showed a notable increase over the individual models’ outputs. The area under the curve increased by over 6% when cross-validation methods and AUC-ROC performance indicators were used, compared to the CNN model alone and the composite model. Performance assessment has shown that adding genetic and imaging data to composite AI greatly improves the precision and resilience of breast cancer detection. This illustrates the revolutionary potential of AI in the early diagnosis of breast cancer and opens the door for more sophisticated diagnostic and personalized treatment options.