Aidoc AI Radiology Foundation Model: 2,000 Hospitals, 60 Million Annual Cases, FDA Clearance for 14 CT Triage Conditions, Radiology AI Enters Scale Era
In 2026, AI radiology diagnostics has evolved from experimental pilots to a core tool within healthcare systems. AI radiology companies like Aidoc are reshaping medical imaging workflows at unprecedented scale. According to the latest data, Aidoc now operates in nearly 2,000 hospitals, processing over 60 million cases annually, and received FDA clearance in January 2026 for a foundation model-powered CT triage solution covering 14 distinct clinical conditions.
Radiology AI's Scale Milestones
Aidoc's Core Metrics
| Metric | Data |
|---|---|
| Hospital Coverage | Nearly 2,000 |
| Annual Cases Processed | Over 60 million |
| FDA-Cleared Algorithms | 50+ |
| Latest FDA Clearance | January 2026 (foundation model CT triage, 14 conditions) |
Industry-Wide Scale
According to radiology business media and FDA data, by end of 2025, the FDA had authorized 1,451 AI-enabled medical devices, with radiology products accounting for 1,104 (76%). An additional 68 radiology AI algorithms received clearance in Q1 2026.
Foundation Model-Powered CT Triage: Technical Breakthrough
Aidoc's January 2026 FDA clearance represents an important technical evolution in radiology AI — shifting from single-condition algorithms to foundation model-powered multi-condition triage.
Limitations of Traditional Approaches:
- Each clinical condition requires independently trained algorithms
- No knowledge sharing between algorithms
- High deployment and maintenance costs
Advantages of the Foundation Model Approach:
- Single foundation model covers 14 CT triage conditions
- Cross-condition knowledge transfer improves overall performance
- Unified deployment and update mechanism
- Higher cost efficiency
14 Covered Conditions include: Pulmonary embolism, intracranial hemorrhage, aortic dissection, pneumothorax, and other acute life-threatening conditions, plus early identification of various chronic diseases.
Radiology AI Market Landscape
Major Players
Large Equipment Manufacturers:
- GE HealthCare, Siemens Healthineers, Philips, and Canon lead in FDA clearance volume
Pure-Play AI Specialists:
- Aidoc: Nearly 2,000 hospitals, 60M annual cases, 50+ cleared algorithms
- Viz.ai: Focused on stroke and vascular triage, 1,700+ hospitals, 50+ cleared algorithms
Regulatory Pathway
95-97% of radiology AI clearances use the 510(k) pathway (substantial equivalence) rather than independent clinical trials. This pathway accelerates market access but has also raised discussions about the adequacy of clinical evidence.
Clinical Integration Challenges and Opportunities
The "Generalizability Gap" Problem
A persistent challenge in 2026 is the "generalizability gap": models may lose up to 24% of their specificity when deployed at institutions different from those used during training.
Mitigation Strategies:
- Multi-center training datasets
- Continuous post-deployment monitoring
- Institution-specific model fine-tuning
Reimbursement Barriers
Despite the large number of FDA-cleared AI tools, reimbursement remains a major hurdle. For common diagnostic tasks like fracture detection or incidental finding identification, AI algorithms are unlikely to receive independent CPT codes, as radiologists are already tasked with identifying these findings as part of standard interpretation.
Evolution of Evaluation Frameworks
As standard benchmarks become saturated, the industry is shifting toward multidimensional evaluation frameworks assessing:
- Safety
- Real-world workflow integration
- Agentic capabilities
- Rather than just accuracy on static datasets
RSNA 2026: The Annual Radiology AI Showcase
From November 29 to December 3, 2026, Chicago will host RSNA 2026 (Radiological Society of North America Annual Meeting), the essential meeting for medical imaging and radiology, featuring an extensive AI Showcase highlighting FDA-cleared, production-grade clinical AI.
RSNA's Key Contributions:
- RSNA Imaging AI Certificate Program v2.0: Updated to integrate clinical use cases into everyday diagnostic workflows
- ATLAS (Annotated Library of AI Systems): Provides model and dataset cards to help researchers evaluate and compare AI tools
- MIRA (Medical Imaging Resource for AI): Research data repository
Asia-Pacific Perspective
The radiology AI market in the Asia-Pacific region is developing rapidly but faces unique challenges:
Market Opportunities
Workforce Shortages: Many Asia-Pacific countries face severe radiologist shortages. Developed markets like Japan, South Korea, and Australia face aging radiologist populations, while emerging markets like China, India, and Southeast Asia have insufficient absolute numbers of radiologists. AI triage tools can significantly improve the efficiency of existing radiologists.
Infrastructure Investment: Asia-Pacific governments are heavily investing in healthcare AI infrastructure. China's "Healthy China 2030" plan, Japan's "Society 5.0" strategy, and South Korea's "Digital New Deal" all prioritize healthcare AI.
Localization Challenges
Data Localization: Asia-Pacific countries have strict restrictions on cross-border transfer of medical data, requiring AI systems to be deployed locally or use local data centers.
Regulatory Differences: Regulatory frameworks vary by country — FDA clearance cannot be directly used in Asia-Pacific and requires approval from local regulatory bodies (China's NMPA, Japan's PMDA, South Korea's MFDS).
Population Diversity: Asia-Pacific's population diversity (different ethnicities, body types, disease spectra) requires AI models to be validated on local data to ensure generalizability.
Conclusion
Aidoc's 2,000 hospitals, 60 million annual cases, and 14-condition FDA-cleared foundation model represent a mature transition in radiology AI from "proof of concept" to "scaled deployment." With the application of foundation model technology, radiology AI is breaking through the limitations of single-condition algorithms, evolving toward broader, more flexible clinical applications. For Asia-Pacific healthcare systems, this trend presents both opportunities and challenges — how to effectively introduce and deploy these technologies within local regulatory requirements and data restrictions will be a core issue for the coming years.
