The following is a summary of “Investigating the Effects of Artificial Intelligence on the Personalization of Breast Cancer Management: A Systematic Study,” published in the July 2024 issue of Oncology by Sohrabei et al.
In the realm of cancer management, ensuring that each patient receives the most appropriate and timely specialized treatment is crucial. Precision oncology, which focuses on tailoring targeted therapy to the specific genetic alterations of patients with breast cancer, aims to not only mitigate disease progression but also enhance patient survival rates. Integrating artificial intelligence (AI) into this paradigm offers significant potential for refining treatment strategies by identifying and selecting the most effective therapeutic options for individual patients.
To explore this integration, the study group conducted a systematic review of relevant literature, drawing from PubMed, Embase, Scopus, and Web of Science databases as of September 2023.
The search employed keywords such as “Breast Cancer,” “Artificial Intelligence,” and “Precision Oncology,” along with their synonyms, specifically targeting articles discussing these intersections. Researchers excluded descriptive, qualitative, review, and non-English studies and assessed the quality and bias of selected articles based on the SJR journal and JBI indices, adhering to the PRISMA2020 guidelines. From this review, 46 studies were identified, highlighting the application of AI in personalized breast cancer management.
About 17 studies employing deep learning techniques demonstrated considerable success in predicting treatment responses and prognoses, thus contributing significantly to personalized treatment approaches, 2 studies utilizing neural networks and clustering methods provided reliable indicators for predicting patient survival and tumor categorization, while another study utilized transfer learning for treatment response prediction. Furthermore, 26 studies utilizing various machine learning techniques showed improvements in breast cancer classification, screening, diagnosis, and prognosis. The most common methods included Naive Bayes (NB), Support Vector Machines (SVM), Random Forest (RF), XGBoost, and Reinforcement Learning, with models achieving an average area under the curve (AUC) of 0.91.
Accuracy, sensitivity, specificity, and precision metrics for these models were reported to range between 90% and 96%. These findings underscore the efficacy of AI in aiding physicians and researchers by revealing complex patterns within omics and genetic data, suggesting that intelligent processing and classification of these data through advanced AI techniques have the potential to revolutionize breast cancer management and enhance patient outcomes significantly.
Source: bmccancer.biomedcentral.com/articles/10.1186/s12885-024-12575-1
