JACR: AI, blockchain dominate future of medical imaging

As technological innovation, new business models and artificial intelligence (AI) have emerged to the forefront of the medical device industry over the past decade, medical imaging in particular has benefited the most from these trends.  

A new analysis conducted by researchers at McKinsey & Company in New York, published Dec. 4 in the Journal of the American College of Radiology identified four emerging technologies that are the focus for global medical imaging startups: AI and machine learning, blockchain, real-time imaging (i.e. three-dimensional (3D) imaging, virtual-reality), and incremental innovation technologies (i.e 3D printing, portable imaging, telemedicine).   

For their study, Alan Alexander, MD, and colleagues analyzed the global startup landscape for new and emerging technologies in data acquisition, visualization, interpretation, storage and sharing in medical imaging. The researchers were left with a total of 146 startups that have received more than $1.8 billion in investments in 446 separate transactions over the past five years. Of these startups, AI and blockchain were two of the most promising technologies found.  

AI and machine learning  

AI and machine learning startups have the most transactions and represent more than 20 percent of companies included in the analysis, according to the researchers. The technologies had the greatest number of startups (32) with more than $500 million in investor transactions, and the greatest revenue, strong investor interest and high growth potential.  

“It [AI] has the potential to transform the way health care is conducted, with AI and machine learning solutions complementing the work of physicians to enable the development of new treatment paradigms,” Alexander et al. wrote. 

As the number of patients and complex cases continues to increase, the researchers explained that AI and machine learning could help radiologists increase their reading efficiency and accuracy, prevent unnecessary testing and more seamlessly integrate imaging data into electronic medical records.  

Of the 32 startups, the researchers found 22 percent of companies were focused on CT, 13 percent on each mammography and MRI, nine percent on ultrasound, three percent on each nuclear imaging and x-rays and 31 percent non-specified.  

Additionally, 72 percent of the AI companies are developing deep-learning technologies to identify lesions or abnormalities; 12 percent include electronic medical record integration for auto-population of measurements and comparisons; nine percent for workflow management to prioritize scans; and six percent for the use of reported findings to improve decision making.  

Blockchain  

Blockchain—an area “worth watching”—could serve as a cheaper, more secure method to transfer real-time medical imaging data between imaging facilities and to patients. It may ultimately make imaging scans more accessible to patients, providers and researchers.  

“By enabling users to store and access data across a distributed ledger, blockchain will facilitate information sharing and support the development of a more integrative approach to medicine,” the researchers wrote, noting 10 percent of the startups analyzed were health-related blockchain companies. 

Like AI and machine learning, the researchers explained blockchain could be of great benefit to radiologists by eliminating unnecessary repeat imaging scans, reducing costs and allowing clinicians to compare scans over time to determine disease progression or identify findings with no clinical significance.  

Other technologies gaining momentum  

Following closely behind AI and blockchain—improved visualization technologies such as 3D, virtual- reality, intraoperative, nuclear and portable imaging—are on the horizon of medical device innovation, according to the researchers.  

A total of 19 startups are focused on 3D imaging and virtual-reality and are among the longest-established companies in the analysis. The companies attract significant funding activity, with the third-highest transaction value after AI and machine learning and intraoperative technologies, according to the researches.  

“With interactive 3D visualization platforms, surgeons and interventional radiologists will no longer have to rely on 2D images from a scan when they make incisions, operate, or perform other interventions,” the researchers wrote. “These technologies can enhance pre-operative and procedural planning, enable greater accuracy in biopsies, and act as a valuable teaching tool in surgery and other interventions.”   

In comparison, intraoperative technologies—which can span across all modalities and often require portable capabilities—can help with planning and execution, particularly for applications requiring extensive data acquisition in real-time. Noninvasive diagnostic techniques such as nuclear medicine can help with cancer diagnosis and management.   

“As nuclear imaging grows and moves away from metabolic imaging for cancer localization to pathology-specific cancer diagnosis (for instance, the combining of ligands specific to prostate or breast cancer with radioisotopes for definitive diagnosis), imaging companies will have increasing opportunities to engage in cancer diagnostics and potentially therapeutics as well,” the researchers wrote.