1. To learn about the disruptive role that big data and artificial intelligence can play in medical imaging.
2. To appreciate that these technologies are key to realise the full potential of precision medicine.
3. To appreciate where the big challenges and opportunities are.
The analysis of big data with artificial intelligence techniques will have an enormous impact on disease prevention, cure and care, and by 2030 it will have dramatically changed the landscape of the health-care system. This session will show examples of possible large benefits of big data analytics in health care from the imaging perspective. Specifically, in this introduction I will provide examples to show that the use of AI and big data should be embraced, as it has a large potential to realizing the potential of precision medicine and precision health, both in dementia and oncology.
1. To understand the importance of creating IT infrastructures for the management of quantitative medical image analysis solutions, both for research and clinical applications.
2. To become familiar with the main storage and computational requirements for high performance infrastructures.
3. To learn about how to move from research to clinical application in the field of quantitative imaging and biomarkers.
Quantitative image analysis solutions and artificial intelligence (AI)-based methods usually find interoperability issues when trying to transfer information with hospital information systems (HIS) and electronic health records (HER). Appropriate quantitative data management would allow for a better understanding of the tissue and organ alterations in the disease and their relationship with patient characteristics and associated clinical or lab data. The image reading process is today in constant improvement, from traditional image reading by the naked eye and free text reporting of the findings, towards the inclusion of new technologies in the loop such as AI computer-aided detection and diagnosis (CAD), imaging biomarker extraction and structured reporting. To cover all clinical needs using AI solutions, quantitative imaging and structured reporting, a technological framework was engineered. The platform is modular and allows the installation both in local and cloud environments. The solution also includes a built-in anonymization module and allows for the incorporation of AI and image processing plug-ins. After the analysis of the images through the different pipelines, a quantitative structured report is generated, which can be integrated with hospital information systems. The platform allows for the mining of quantitative data extracted from patients, a functionality which is not available nowadays in PACS systems and workstations.
1. To learn about machine learning concepts used for diagnosis, prognosis, and quantitative analysis of medical imaging data.
2. To understand the challenges to introduce these techniques in clinical practice.
3. To appreciate the potential of machine learning in radiology.
This lecture introduces the current state-of-the-art techniques in machine learning in medical imaging, highlights the potential of these techniques for computer-aided diagnosis and quantitative imaging biomarkers, addresses challenges related to variations in scan protocols and a requirement of annotated data for model training, and presents examples of successful applications of machine learning in radiology.