Big Data Will Change Hospital Structures
Big data evaluated by artificial intelligence will disrupt medicine, including radiology. “By 2030, it will have dramatically changed the landscape of the healthcare system,” said Biomedical Image Analysis expert Wiro Niessen at ECR 2018.
“Medical imaging has already adopted this field, because we want precision medicine,” said Wiro Niessen, Erasmus Medical College Rotterdam, The Netherlands.
Niessen believes that AI and big data should be further embraced, due to their large potential to realize precision medicine and precision health. The field will help to individualize decisions about when to do an intervention, it will predict therapy success and support the development of preventive strategies.
Niessen addressed dementia as one example. “Large scale data analytics in longitudinal population neuroimaging studies, especially when combining imaging with other clinical, biomedical and genetic data, provides a unique angle to study the brain, both in normal ageing and disease,“ he said.
Big data pools like the Rotterdam Study that has been collecting data from more than 15,000 participants seen and imaged every four years since 1990 provide a basis for AI analysis. A library of quantitative imaging parameters is part of the process. “We have started to compare individual measurements to reference data to get objective results and build predictive models,” he said.
Niessen identified “radiomics”, i.e. the connection of imaging and genetic data as another important field for radiology. This approach can be used to improve tumor characterization and select the right therapy. The new WHO classification for cerebral tumors already relies on this stratification, with glioma grading highly depending on 1p/19q mutation presence.
“We combine information that will altogether influence clinical decision making,” stated Niessen. This will disrupt processes and structures. More cooperation and completely new hospitals structures will likely result from this.
Niessen believes human intelligence to be complimentary to AI. “In the end we will have medical imaging instead of data science,” he concluded.