Artificial Intelligence and Radiomics in Cardiothoracic Imaging: Current Applications and Future Trends
- Artificial Intelligence (AI) research has been performed within the fields of mathematics and computer science since at least the 1950s. [1,2] Originally, much of AI research aimed to solve relatively straightforward problems that were intellectually diﬃcult and time consuming for human beings. Over time, AI research has evolved to focus on developing algorithms for solving pattern recognition problems that may be relatively easy for humans to perform but are difficult to explicitly describe (e.g.
- “It is the framework which changes with each new technology and not just the picture within the frame.” Marshall McLuhan
- The capabilities of artificial intelligence (AI) are rapidly progressing and the research community is getting increasingly interested in its possibilities. With advances in computing power, storage capabilities and innovative algorithms, the use of machine learning (ML) and deep learning (DL) in medical imaging research has grown rapidly. The public availability and open source nature of AI libraries, such as Tensorflow  and PyTorch , make research and use of AI algorithms possible not only for professional computer scientists but also for clinical researchers from a wide variety of fields.
- The current radiological practice is facing profound changes with the progressive implementation of artificial intelligence (AI) in medicine , empowered by increasing computational capacity, availability of healthcare data and advanced image analysis techniques.
- Artificial intelligence has become a hot topic in radiology these last years, with already 150 deep learning articles only focusing on medical imaging in 2018 . Machine learning gives computers the ability to learn from data and reproduce human interpretations without being explicitly programmed. Computer vision, a scientific field of particular interest for radiologists, shares a number of objectives with machine learning. The goal is to make it possible for machines to analyze, process and understand digital to automate tasks that the human visual system can do.
- As in the case of most disruptive technologies, assessment of and consensus on the possible ethical pitfalls lag. New AI applications and start-up companies seem to emerge daily. At the start of 2019, funding in imaging AI companies exceeded $1.2 billion . Yet, questions of algorithm validation, interoperability, translation of bias, security, and patient privacy protections abound.