From Big Data to Bedside (BD2B): Towards AI-Based Precision Oncology
Cancer is mainly caused by somatic genome alterations that perturb cellular signaling pathways, and it is anticipated that precisely targeting patient-specific genomic alterations of individual tumors (precision oncology) will lead to more effective therapies. However, while genome-scale data from an individual patient can be readily obtained to guide contemporary molecularly targeted therapies, only a small fraction (< 10%) of all patients benefit from current precision oncology approach, and majority of patient are treated by standard of cares following current guide lines, which are not personalized. In this presentation, I will discuss different artificial intelligence technologies that can advance precision oncology, including causal inference methods for revealing the disease mechanisms of each individual tumor, causal network methods for discovering cancer pathways, and deep learning methods to infer the state of signaling machinery of tumor cells. I will further discuss how information derived from such analyses can be used to guide personalized application of anti-cancer drugs and present our results from pre-clinical studies.