CIOs must evaluate best use cases for AI in medical imaging

Medical imaging plays a big role in helping physicians care for their patients, but the systems that are responsible for managing medical images have traditionally focused on storage and retrieval. However, healthcare organizations are seeing that medical imaging systems such as PACS and VNAs offer advanced capabilities like analytics and AI that can change the way they interact with these products.

Hospital CIOs don't always view the implementation of PACS or VNAs as a complex endeavor. Most imaging platforms share the same DICOM standard and image storage and retrieval features, but rolling out a new PACS or VNA requires a comprehensive infrastructure and the buy-in of different departments regardless of which product is selected. Some of the common features across the different imaging platforms include:

• image storage, export and routing
• quick access to images from multiple viewing stations and devices
• support for DICOM standards
• report generation
• user templates for patient studies
• HL7 support

In recent years, medical imaging has seen some major changes and a race by vendors to include AI capabilities into their imaging systems. This is the result of advancements in image processing, machine learning and algorithms that allow computer software to analyze large numbers of images and learn from them. This allows the system to detect with a high level of accuracy things that a human may not always be able to see.

A recent example of the use of AI in medical imaging is the image-based diagnosis of diabetic retinopathy. In this scenario, the use of AI alongside images captured by eye imaging equipment is able to determine whether the patient suffers from the condition. The procedure takes a few minutes and can be performed at a physician's office through the IDx-DR device, and within roughly 20 seconds of the image capture, the system can provide the binary result -- either a yes or no -- to the provider and patient. This procedure previously required a specialist to perform the test. In some cases, the condition went undetected and caused more complications for the patient.

For hospitals, the use of AI in medical imaging provides them with an opportunity to take advantage of new capabilities that can enhance patient care and provide new efficiencies and improve productivity. Some of these capabilities include:

• abnormality detection through machine learning and image processing
• smart dictation for note taking using natural language processing
• health data mining
• image processing and analysis
• high-risk patient detection

The transition to intelligent PACS or VNAs will require proper planning on the infrastructure and the end-user fronts. CIOs will need to ensure that their IT environments are able to meet the different requirements that the systems will need, like storage, processing power and integration support with the EHR. As for how to prepare end users for the new PACS systems with AI capabilities, more training and process reengineering will likely be the main focus of the change. Because AI will require more data, integration with other systems may be a requirement as well.

Despite the increase in approvals seen in hospitals when considering medical imaging systems that include AI, many of the products in the market today are still in their early stages. More will need to be done to evaluate the different use cases for AI in medical imaging, but it's clear the technology is here to stay.