1. To explain the likely future development of structured reporting.
2. To highlight the opportunities which will arise from availability of radiological data in formats which can be searched and used as imaging biobanks.
3. To discuss the data protection issues which arise from these developments.
As artificial intelligence increases its penetration into all aspects of life, radiology will change in the future. Radiology reports will become more standardised in format and language, and will become rich sources of information for teaching, research and machine learning, feeding back into a loop of continuing improvement of technological tools. This session will explore how these processes will operate, how radiology report data will be mined and utilised in the future, and what issues this will highlight in terms of data ownership and protection.
1. To discuss the role that the imaging biobanks can play in providing the data necessary for machine learning and for the assessment of the performances of diagnostic tools.
2. To understand the role ontologies play in the standardisation of image annotations in imaging biobanks.
3. To understand the need to collaborate with the domain of (specimen) biobanking.
Significant progress in computer-aided or even automated image interpretation are expected in the near future. Machine learning and assessment of the performances of these new tools will require large quantities of annotated image data.
The development of imaging biobanks should be able to provide such data. The presentation will summerize the challenges of gathering large sets of image data, either from past clinical research studies, or from large cohorts and population studies. Particular emphasis will be put on the definition of consensual ontological models to guarantee a good standardisation of semantics borne by image annotations. Indeed, the latter must cover a wide scope from general characteristics of the clinical cases (e.g. known pathology, patient sex and age) up to quantitative image biomarkers derived from the images. The issue of relating image data to other biological data (e.g. genetic / genomic) will be underlined, justifying a strategy of designing information models for imaging biobanks that are consistent with those used in biobanking (e.g. BBMRI-ERIC).
1. To discuss the data requirements for artificial intelligence and big data.
2. To understand how structured reporting can facilitate research in machine learning.
3. To become familiar with informatics standards, relevant to structured reporting.
There has been a tremendous increase in interest towards artificial intelligence and machine learning applications in radiology. In order for such systems and algorithms to perform sufficiently well, large amounts of training data are required. These data need to be either standardized and of adequate quality or available in large enough amounts, that enough relevant information can be extracted. However, in current clinical and radiological IT ecosystems access to relevant pieces of information is difficult. This is mostly due to the fact, that a large portion of information is handled as a collection of narrative texts and interoperability is still lacking.
This lecture will give an overview on how structured reporting can help to facilitate research in artificial intelligence and in the context of big data. Potential approaches to implementing structured reporting will be discussed, focusing also on interoperability and standardization.
1. To understand the current and future role of data.
2. To understand the issues surrounding data control and ownership.
3. To discuss the role of AI in the future of radiology.
Recent years has shown an explosive growth in the use of Artificial Intelligence (AI) and Deep Learning not in the least for medical applications. These new technological developments have started a whole new discussion on privacy of data processed by these computerized systems. Especially those applications in health care demand a high level of patient privacy and patient data security. The actual ownership of medical data is also part of this discussion where we can have different situations with original, de-identified, anonymized and processed data. Questions like what data is still personal data for an individual patient or participant in a clinical trial and who actually owns the data that is produced with self-learning computer systems. This educational course will discuss these issues with respect to ownership, intellectual property and the different aspects that are involved in this when moving towards the era of AI.