International Comparisons: Pooling Patient Data to Evaluate Longterm Use Drugs for Its Carcinogenicity and Repositioning Using Artificial Intelligence Techniques

Cancer, the most challenging health problem globally. Annually 14.1 million new cancer cases and 8.2 million cancer-caused deaths occurred worldwide. [1] Costs of cancer therapeutics and supportive care drugs keep growing and reached $107 billion globally in 2015.[2] The financial burdens on patients becomes heavy, e.g. in US the average costs per patient in first year after diagnosis with breast cancer are higher than $60,000 even for disease at stage 0.Every year more than 5,000 people died of cancer in Taiwan. Cancer patients in hospitals ranked fourth in 2008, and new patients accounted for 1.5% in Taiwan. Overall survival rate was 65.6% and 10 years was 44.7%.[4] Cultural factors result in higher than usual usage rate of healthcare by Taiwanese citizens. Today each person in Taiwan visits a health facility on average 14-15 times annually. For each visit patients gets prescribed to 4-5 medications. The carcinogenic effect of drugs that prescribed for chronic disease management still remains unclear. Several drugs, for example, aspirin [5-7] and metformin [8, 9] were considered to help reduce cancer, while certain benzodiazepines were found to increase the cancer risk [10, 11]. Statins which considered as Long term use drug (LTDs) seems remain controversial and inconclusive [12-14], as different statins show their own preferences on cancer types.Nowadays, artificial intelligence will determine the future of medicine. Although AI has been around for decades, new advances have ignited a boom in deep learning. The AI technique powers self-driving cars, super-human image recognition, and life-changing events with advances in medicine. AI can help doctors make faster, more accurate diagnoses. It can predict the risk of a disease in time to prevent it and help researchers to understand how genetic variations lead to disease.The overall objective of this project is to build an intelligent international generalizable model that can estimate personalized liver cancer risk based on diseases history and prophylactic drug intervention changes. These models are expected to reach a certain level of precision based on patient-driven big health data without asking extra examinations using large-scale health data and the aid of deep learning methods. The methods will include a nationwide health insurance database (NHIRD) and electronic health records (EHR) from five TMU affiliated medical institutions and international data repositories from the Observational Health Data Sciences and Informatics (OHDSI) [15] International consortium collaborator, Canadian and MIT-Harvard University using their ICUMIMIC Database.This project will follow a multidisciplinary international approach using Artificial intelligence models and informatics working TMU in close collaboration and discussion with OHDSI consortium such as Canada, US, Japan, Korea, Australia and Taiwan to develop an international generalizable liver cancer prediction model besides Taiwanese population.One of the many goals of this project is to be able to compare the model across different populations for its generalizability and new knowledge regarding the effects of cancer aids in various levels of health and disease, promoting thus scientific excellence regarding the effects of cancer in the domains of toxicology and physiology via state of the art approaches of international multidisciplinary consortium.