svedoliver - Fotolia


Radiology to gain from artificial intelligence in healthcare

Some radiologists may be unsure of how much they can trust artificial intelligence, but they should see it as a tool to review their work and check it against similar cases.

Artificial intelligence and natural language processing are making their way into the medical imaging field, and radiologists have mixed feelings on these developments because they might encroach on the invaluable human element of evaluating diagnostic images. The progression of artificial intelligence can be seen in virtual assistants and self-driving cars, while in healthcare it can be used to assist in the diagnosis and care of a patient.

Radiological images are complex, and it requires a human ability to interpret and detect patterns in patients' CT scans, MRIs and X-rays. Much of the analysis performed on medical images is learned over years of training and experience. But by combining advancements in image processing with artificial intelligence in healthcare, there is an opportunity for technology to supplement the work of radiologists and help them with their duties.

There are some artificial intelligence products that are presently available to hospitals and radiologists to assist healthcare professionals with various aspects of their work. Automation and artificial intelligence in healthcare can be directed to close human performance gaps in some of the following areas.

Where artificial intelligence in healthcare can be applied

Volume of studies
The amount of captured and stored medical images is increasing. As the quantity of images has gone up, so too has the amount of time it takes radiologists to review the data. Artificial intelligence in healthcare can take some of this work away from radiologists by processing images and scanning medical studies to quickly detect patterns or abnormalities that could be missed by the naked eye. The artificial intelligence system could then pass the results of its review to a radiologist for confirmation.

Reporting and classification
Natural language processing is another technology radiologists can use to assist them with documentation and reporting. A natural language processing system can read reports from different sources and assist in classifying the data based on its content.

Deeper image reading
Advanced medical image processing research done by Microsoft developed algorithms that can detect irregularities in patient scans that aren't always caught by the human eye. These capabilities and technology have been successfully used to aid in mammography and other medical specialties.

One pixel VIDEO: Are smart machines paving the way
to a safer workplace?

Big data
The large volume of medical imaging data available in hospital's data centers presents a great opportunity for advanced analytics and data mining. A system that combines the power of image processing and medical artificial intelligence can catalog and scan millions of images to find those with similar appearances. Microsoft's research project InnerEye is an example of a system that can analyze medical images for signs of diseases or other anomalies.

Cost reduction
There are many vendors that sell artificial intelligence products and services. Some of these services, such as those offered by Enlitic, Inc., promise to reduce costs, while boosting the accuracy and speed of workflows for radiologists and other healthcare professionals.

Artificial intelligence is bound to play a more significant role in the near future of healthcare. With its ability to evaluate large data sets and detect insights faster than any human can, it is becoming the tool that every physician will want to have as an assistant. While its availability is currently limited, medical artificial intelligence will become part of EHRs and other applications used by physicians in years to come.

Next Steps

3D printing helps during planning phases of surgery

A comparison of vendor neutral archives and picture archiving and communication systems

Radiology at the center of the largest North American healthcare tech event

How home virtual assistants can increase patient engagement

Dig Deeper on Clinical decision support systems