Will Artificial Intelligence Replace Radiologists One Day?

Will Artificial Intelligence Replace Radiologists One Day?


Artificial intelligence (AI) has undergone dramatic growth in recent years; its effects can be seen in virtually every field, from oil drilling to games of chess. Deep learning, a method which teaches computers to recognize objects within pictures, has significant potential impact on the medical imaging field.


That’s why some AI researchers say the technology could spell the end of human radiologists. Computers, they say, would be able to quickly, cheaply, and accurately perform the same job functions. While the power of AI cannot be denied, it seems premature to make assumptions about the important role trained human radiologists play in the contemporary health care industry. Artificial intelligence will almost certainly help radiologists do their jobs; we just doubt it can replace them entirely.


Radiologists may certainly be worried about the future of their profession, but any concern would be premature at this point. In fact, the adaptation of AI into imaging technology has great potential for increased productivity among radiologists and increased profitability for those institutions housing them. There are several current and future avenues in which AI may prove a useful augmentation to radiologists, as well as certain restrictions which ensure radiologists will continue to have gainful employment well into the future.


How AI Can Improve Radiology as the Tool of Human Radiologists


The first way in which AI may serve radiologists is in a quicker diagnosis of certain types of fractures and cancers. This would free the radiologist for more difficult-to-detect cases that could not be handled by current machine learning technology. In many cases, AI could also confirm a diagnosis that may have otherwise been in doubt, providing clinical decision support. The increased usage of deep learning in imaging also would allow the radiologist to spend more time focusing on the patient and their treatment plan, potentially creating greater job satisfaction for that individual.


A second major reason for the medical field to embrace AI is the potential for increased demand in imaging procedures. In countless cases throughout history, increased automation meant greater efficiency, which in turn brought the price of the given product down. Often throughout history, increased automation brought about fears of workers losing their jobs; however, the lower prices usually resulted in such a significant uptick in consumer demand that more jobs were created than lost. The same potential exists for the relationship between AI and radiologists.


Third, many conflate the type of AI learning being applied to radiology with the AI seen in science fiction movies. The deep learning being applied in to diagnose fractures or occurrences of cancer is an extremely focused technology used for that purpose and that purpose only. The AI is not suggesting a treatment pattern, or consoling a patient who has received devastating news. AI is only reading an object in a picture, albeit to a highly accurate degree in some cases.  


Why Human Radiologists Will Always Be Part of Diagnostic Imaging


No machine will ever be able to replace the empathy that medical providers develop throughout their careers. There’s no substitute for a warm hug, a meaningful conversation, or time spent with a trusted physician.


Artificial intelligence has the potential to greatly alter the field of radiology in many ways that generate a positive impact for all those involved, from the patient to the radiologist to the medical institution. Imaging specialists need not worry about job loss at the hands of a faceless algorithm. Instead, opportunities to increase patient well-being and diagnoses accuracy should be adopted whenever possible.      


Advances in Machine Learning (AI) for MRI scans

Open-Source MRI Dataset from USC Now Available to Researchers


The University of Southern California has released an open-source dataset of anatomical brain images taken from MRIs of stroke victims. The dataset is intended to spur advances in machine learning by providing a large set of manually-traced lesions.


Manually-traced lesions are useful, but labor intensive.


Researchers have attempted to automate lesion segmentation through algorithms. However, this automation is in its primitive stages, and machines cannot yet identify lesions with great accuracy. Thus, manually-traced lesions are still the gold standard, but require a large amount of work from a trained neuroanatomy expert.


USC’s dataset attempts to bridge the gap between human tracers and machines. By providing 304 T1-weighted MRIs with lesions segmented by a human, the study’s authors hope computer programmers can develop an accurate lesion segmentation algorithm. The dataset is available for download free of charge here.


Strokes are a leading cause of death and disability in the United States.


Mortality rates from strokes have steadily declined worldwide, but around two thirds of stroke survivors suffer long-term disabilities that affect their daily activities. This situation has led scientists to focus on what interventions provide the best outcomes for stroke survivors.


Doctors have opportunities for intervention at both the acute and chronic stages. In the former, intervention can save neural tissue and promote functional recovery. In the latter, rehabilitation can help long-term recovery.


Magnetic resonance imaging can aid doctors in making intervention decisions.


Clinical brain images taken within 24 hours of a stroke help doctors determine whether to administer thrombolytic drugs or perform surgery to save neural tissue. Because clinical scans are taken for almost all stroke victims, there have been great strides in using large-scale datasets of these acute scans for predictive modeling.


Unfortunately, sub-acute and chronic scans are given less and therefore harder to obtain, making predictions at these levels less advanced. That’s one thing the study’s co-author Sook-Lei Liew would like to change.


“The goal of ATLAS is to generate a dataset that machine learning and computer scientists could use to develop better automated algorithms to identify the lesions,” Liew told Health Data Management.


Machine learning requires large, accurate datasets to train and to test.


Liew hopes that computers will eventually be able to identify biomarkers in stroke patients, making it easier to prescribe the appropriate rehabilitation therapy and treatment. Her next step is to create a separate dataset used to test the algorithms developed using the current dataset.


“In machine learning, you always need a training dataset and a testing dataset.” Liew noted. “Even if people aren’t interested in stroke, it’s also an interesting dataset to train any sort of computer vision algorithm because it’s a challenging problem.”




Slabodkin, G. USC Releases MRI Stroke Dataset To Spur AI Research. Health Data Management. Available here. Accessed February 22, 2018.  


Liew, S et al. A Large, Open Source Dataset of Stroke Anatomical Brain Images and Manual Lesion Segmentations. Scientific Data. February 20;(5):180011. doi:10.1038/sdata.2018.11

Accessed February 22, 2018.   


Feigin, VL et al. Global and regional Burden of Stroke During 1990-2010: Findings From the Global Burden of Disease study 2010. The Lancet. January 2014;(383)9913:245-255. doi:10.1016/S0140-6736(13)61953-4 Accessed February 22, 2018.