AI adoption challenges The emergence of AI comes with deep challenges, particularly for the medical facilities and radiologists themselves who will interact most closely with these new technologies.
Hospitals will see their financial and IT sides impacted as they manage multiple AI vendors. The complex hospital mechanism involving medical professionals, IT departments and financial functions will need to adapt to the need to evaluate, integrate, and manage many coexisting solutions.
Radiologists will be similarly challenged in their day-to-day work schedules, faced with new applications that include different interfaces and detection schemes. Any effective solution will require behavioral changes: creating new workflows, increasing trust in technology's ability to perform with or without supervision, and resolving conflicts between human interpretation and CDS results.
New models The aforementioned conditions will spur the development of new consolidated models:
App store-type platforms (e.g., Envoy AI) are one alternative, where providers will only have to manage a single technical interface, with the platform itself managing all sub-vendors. These platforms will ideally create multiple integrations for workflow products.
Radiology-as-a-Service (i.e. private practices or teleradiology providers) is an alternative model particularly relevant for smaller providers. This model would provide the added benefits of load balancing and access to specialty radiology, thus offering end-to-end alternatives to in-house radiology clinics.
Lastly, we can envision a model based on bundling of applications. This can take the form of layers on top of existing PACS solutions or of application suites, allowing easy access to multiple applications that together offer solutions for healthcare providers and physicians.
Who will come out on top? As the imaging market evolves, several players will emerge:
Entrenched players from within the current value chain: The PACS, reporting companies, and even scanner manufacturers will potentially make plays to become AI aggregators. They currently own parts of the radiology workflow and can easily embed AI in their products through joint APIs, and are thus well-positioned to provide solution suites.
AI application startups are also strong contenders to become market leaders by either developing broad in-house applications, or by integrating other startups' applications. Startups' competitive advantage is that they begin from the application side and thus grow organically in the organization – a significant benefit on the user side as well. But with limited resources to dedicate to expansion, startups will face some key hurdles.
