Scientific literature is arguably the most objective, comprehensive and timely data source of medical knowledge discovery and evidence based medicine, as the significant clinical outcomes are often reported in a timely fashion in scientific journals. In this project, we proposed a text mining method for acquiring treatment strategies of intrahepatic bile duct cancer from PubMed abstracts. Treatment related concepts including therapeutic procedures, medications, genes and medical devices were systematically extracted and analyzed. A term co-occurrence network was generated to identify the main topics of the latest research on treatment of intrahepatic bile duct cancer. The treatment relations about 4 groups of different concepts, including Therapeutic Procedure, Medication, Care Activity, and Medical Device, were identified and demonstrated. Our study showed that the automatic treatment strategies extraction from literature with a small amount of physicians' verification added is a feasible solution to quickly characterize the overall treatment research progress.