The use of clinical data is key to the continuous improvement of health care and also to accelerate research directed towards prevention, diagnosis, and treatment innovation. At IPO-Porto, healthcare professionals and researchers have the support of several departments that are able to provide relevant data to answer their clinical and scientific questions, while preserving patients’ privacy. Unfortunately, information about the previous medical history and some follow-up data are not available in easily accessible formats, because the registration of these data is not stored in structured formats, existing in .pdf files containing free text. This gap represents an important obstacle to perform retrospective cohort studies and to plan prospective observational or interventional protocols. The aim of this work is to create and validate text mining algorithms to extract relevant clinical data from .pdf files (such as the hospital discharge summaries and other medical reports) in a reliable, safe and confidential way, transforming them into structured format data. This study will only include data from patients with Acute Myeloid Leukemia.