Reviewing the relationship between machines and radiology: the application of artificial intelligence

Abstract

Background The scope and productivity of artificial intelligence applications in health science and medicine, particularly in medical imaging, are rapidly progressing, with relatively recent developments in big data and deep learning and increasingly powerful computer algorithms. Accordingly, there are a number of opportunities and challenges for the radiological community. Purpose To provide review on the challenges and barriers experienced in diagnostic radiology on the basis of the key clinical applications of machine learning techniques. Material and Methods Studies published in 2010–2019 were selected that report on the efficacy of machine learning models. A single contingency table was selected for each study to report the highest accuracy of radiology professionals and machine learning algorithms, and a meta-analysis of studies was conducted based on contingency tables. Results The specificity for all the deep learning models ranged from 39% to 100%, whereas sensitivity ranged from 85% to 100%. The pooled sensitivity and specificity were 89% and 85% for the deep learning algorithms for detecting abnormalities compared to 75% and 91% for radiology experts, respectively. The pooled specificity and sensitivity for comparison between radiology professionals and deep learning algorithms were 91% and 81% for deep learning models and 85% and 73% for radiology professionals (p < 0.000), respectively. The pooled sensitivity detection was 82% for health-care professionals and 83% for deep learning algorithms (p < 0.005). Conclusion Radiomic information extracted through machine learning programs form images that may not be discernible through visual examination, thus may improve the prognostic and diagnostic value of data sets.

The Artificial Intelligence in Digital Radiology: Part 1: The Challenges, Acceptance and Consensus

“…The challenges in Digital Radiology were also mentioned in the study by Ahmad [ 17 ], in the study by Hamed et al [ 18 ], and in the study by Kottler [ 19 ]. Particular emphasis was dedicated to the challenges in the engineering and in the machine in terms of hardware, software, and impact on the workflow.…”

“…In line with the first objective of the study, we analysed the main challenges in the design and integration of Artificial Intelligence in digital radiology. The eligibility process led to the choice of 20 papers [ 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 ], among which there are mainly reviews (15 in number) (as it could be expected considering the broad topics covered) but also 5 recent scientific articles/focus articles on very specific aspects [ 15 , 26 , 27 , 29 , 30 ].…”

“…Particular emphasis was dedicated to the challenges in the engineering and in the machine in terms of hardware, software, and impact on the workflow. These studies [ 17 , 18 , 19 ] also reported important new activities such as selecting Artificial Intelligence products and vendors; piloting vendors’ Artificial Intelligence algorithms; creating our own Artificial Intelligence algorithms; implementing, optimizing, and maintaining these algorithms.…”

“…A second topic was the revolution of the workflow [ 16 ] of the radiologist; many activities could be possible in relation to imaging and many other activities in relation to the administration of other phases of the patient management process (i.e., patient scheduling). The third topic related to the new spreading IT tools that could be strategic in the health domain [ 17 , 18 , 19 ]. This determined (and hence the fourth topic) the need to intensify the collaboration between [ 19 , 20 , 21 ] the insiders.…”

“…In subsequent studies, it will be necessary to deepen this theme based on the categorization proposed by this study. This categorizing certainly also had an impact on IT tools and training databases [ 17 , 18 , 19 , 28 ], which could be appropriately addressed with further specific studies.…”