By Elad Walach and Dr. Yoni Goldwasser
The transformative impact of AI on healthcare has stopped being a point for debate. It's not a question of "if," but of "how," and "how fast."
The field of medical imaging is a prime example of one of the many healthcare subfields that will feel the effects of AI on their workflow in the near future. AI will bring immense gain at each stage in the imaging value chain, which will no doubt be followed by the challenge of adoption for hospitals and radiologists. AI-focused startups and multinationals alike are both seizing the opportunities presented by this growing area of innovation.
AI will bring value at various points in the healthcare value chain
The imaging value chain can be broken down into several stages, with AI contributing in each:
- Scheduling, administration, patient management, and workflow optimization. Given the current inefficiencies utilizing imaging technologies, the complex interface between various providers and the changing regulatory environment regarding these applications, AI offers a much-needed way of optimizing patient management. Companies like HealthLevel are trying to help radiologists improve efficiencies by providing BI and clinical metrics. Other solutions from the HIS/RIS space will continue to come into play in the coming years.
- Pre-scan (e.g., patient positioning): While choosing the correct protocols and ensuring proper patient positioning is ostensibly the responsibility of physicians and technicians, AI algorithms can help prevent errors, improper care, and other difficulties. Bay Labs and Butterly iQ, for instance, use AI to reduce operator dependence in ultrasounds.
- In-scan: One study's results often lead to further studies, wasting resources and prolonging time to diagnosis and care. Through live image processing, AI algorithms could help predict the need to employ new protocols or conduct further studies.
- Post-scan/interpretation: Here is where AI's potential to streamline workflows is particularly valuable. AI can help radiologists prioritize caseloads – reducing, in some settings, more than 90% of diagnosis time for time-sensitive cases. Some AI companies try to target a broad set of clinical use cases (e.g., Aidoc, Zebra Medical, etc.), while others offer deep specialty around specific solutions.
- Predictive analytics/biomarkers – Companies like Quantib and IcoMetrix are trying to find new biomarkers for complex cases like Alzheimer's, helping radiologists spot patterns invisible to the naked eye.