- The market for artificial intelligence (AI) tools to process and analyze medical imaging studies is slated to push past $2 billion by 2023, according to a new report from Signify Research.
Healthcare organizations are increasingly eager to implement machine learning, deep learning, and other detailed pattern recognition algorithms that can provide clinical decision support while improving the efficiency of radiologists, pathologists, and other image-based diagnosticians.
As clinical and financial use cases for AI become more enticing – and fears over the role of artificial intelligence in medicine start to abate – healthcare providers appear somewhat more willing to experiment with new types of data-driven decision-making, the report says.
A previous survey by KLAS research, conducted earlier in 2018, found that 30 percent of organizations are in the process of making plans to adopt AI-driven imaging technologies, while another 17 percent are currently live or piloting new initiatives.
While some wariness over accuracy, reliability, and liability issues persists, healthcare organizations appear to be cautiously optimistic about the potential of AI tools.
“It is becoming increasingly clear that AI will transform the diagnostic imaging industry, both in terms of enhanced productivity, increased diagnostic accuracy, more personalized treatment planning, and ultimately, improved clinical outcomes,” writes report author Simon Harris.
“AI will play a key role in enabling radiology departments to cope with the ever-increasing volume of diagnostic imaging procedures, despite the chronic shortage of radiologists in many countries.”
Source: Signify Research
Deep learning will be a particularly important strategy for AI developers, and will account for more than half of the anticipated market growth over the next five years.
Deep learning methods, such as neural networks, are inspired by the multi-layered data processing structures of human brains, and have already proved their ability to match or exceed the accuracy of human experts when analyzing images.
These pattern recognition tools can identify variations in individual pixels to better delineate the borders of tumors or pinpoint other abnormalities.
Developers and providers are looking to leverage these capabilities in a number of diagnostic regions, and anticipate a relatively broad range of clinical applications. Close to a quarter of the market will focus on neurology imaging, while 21 percent will hone in on cardiovascular issues.
By 2023, fifteen percent of the market will be dedicated to breast imaging, and 14 percent to pulmonary issues. Twenty percent of algorithms are likely to address other parts of the body, including orthopedic concerns.
Source: Signify Research
Harris notes that a number of barriers may still stand in the way of widespread adoption. Regulators are still working to develop guidelines and rules for the nascent industry, which can be difficult in the absence of large-scale validation studies that prove the reliability and trustworthiness of emerging algorithms.
Technical design and implement challenges also abound, with radiologists remaining deeply concerned about how new clinical decision support tools can be integrated seamlessly into their workflows.
Providers may also struggle with implementing artificial intelligence tools that work harmoniously with their legacy systems.
“Healthcare providers are reluctant to purchase AI tools from multiple software developers due to the vendor specific integration and implementation challenges and the administrative overhead,” the report points out.
“Algorithm developers need to establish effective routes to market, such as distribution deals with the established medical imaging vendors and the new breed of vendor-neutral AI platforms.”
Provider organizations are likely to have a large number of options to choose from during the next few years, and may be able to take advantage of offerings from larger, well-established companies in addition to the myriad start-ups promoting innovative solutions.
An increasing emphasis on standards-based technologies and application programming interfaces (APIs) may help the industry overcome the bottlenecks created by legacy systems and proprietary data formats.
As medical imaging analytics tools become more readily available, providers are likely to find strong incentives to investigate the best way to integrate artificial intelligence into their imaging strategies.
Robust growth and increasingly accurate algorithms offer early adopters the chance to get ahead of their peers and augment their workflows effectively with a new generation of highly sophisticated decision-making tools.