Will ML Replace Radiologists?
Declan O'Regan


In this presentation the current state of the art of Machine Learning in medical image analysis will be discussed and what impact this may have on the role of clinical radiologists.


Machine learning (ML) is gaining traction in medical imaging for risk stratification, diagnosis and prediction. Could ML replace some or all of the roles currently performed by radiologists leading to more individualised care, reduced costs and fewer errors? In this presentation the current state of the art in ML image processing will be explored and how it may impact on the traditional role of radiologists.

A brief background to ML will be provided including a primer on Deep Learning techniques and how they are applied to radiology. Examples will be given of cutting edge ML technology in the domain of medical imaging with a critical appraisal of their performance and potential impact on clinical practice. Specific attention will be given to the role of ML in cardiovascular imaging and its potential for novel approaches to disease classification and risk-stratification.

The obstacles to wider development and implementation of ML technology will be discussed including access to high-quality training datasets. Standards for reporting the performance of ML algorithms in radiology will also be proposed.

Lastly, the future of radiology will be discussed in a world increasingly influenced by the drive towards artificial intelligence in healthcare.


No acknowledgement found.



Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)