AI and Medicine: From medical text, medical imaging… to genomics

While you read headlines of Artificial Intelligence and Deep Learning beating doctors in diagnosing diseases such as cancer and interpreting radiographies, you may wonder how far are we from seeing AI actually used and deployed in hospitals.

In this talk I’ll share my journey as a physician and AI engineer through the state of the art in AI applied to medicine:

1. Extracting information suitable for computers from unstructured medical text in cancer clinical trials in English as well as information from chest radiographies clinical reports in Spanish… and building the biggest dataset (patient-wise as of July 2019) in the process.

2. Building deep learning models to classify radiographies across diagnostics, findings, and locations; including a graphical interpretation of where the model is focusing to make the predictions.

3. Exploring how AI is being applied to genetics, specifically in predicting drug-response of patients with certain cancer mutations.

…and the most important of all is how the above is relevant for patient care:

– how to maximize its use so it benefits patients

– how to control biases and confouding factors – how the medical profession is expected to change

– what do we -as physicians- need to use these new advantages

After the talk you’ll have a better understanding of where we are in AI applied to clinical practice and the difference from hype to reality.