RadNet’s Wholly-Owned Subsidiary, DeepHealth, to Use CARPL.ai's Platform to Develop a New AI Control System for Clinical AI Performance and Safety

In This Article:

RadNet, Inc.
RadNet, Inc.
  • DeepHealth and CARPL.ai have established a strategic collaboration to create a unique Artificial Intelligence (AI) control system for image interpretation to ensure AI scalability, performance monitoring, and safety, with the aim to accelerate the adoption of AI. 

  • DeepHealth currently monitors the performance of DeepHealth’s SmartMammo™ AI-powered solution for breast cancer detection at RadNet. Through the collaboration, the two companies aim to expand, productize and scale this control system across more applications to other customers.

  • Furthermore, DeepHealth will embed CARPL.ai’s cutting-edge AI orchestration capabilities that enable easy selection, implementation, and monitoring of appropriate AI models within DeepHealth’s cloud-native operating system, DeepHealth OS.

LOS ANGELES and SOMERVILLE, Mass., Dec. 01, 2024 (GLOBE NEWSWIRE) -- DeepHealth, Inc., a global leader in AI-powered health informatics and a wholly-owned subsidiary of RadNet, Inc. (NASDAQ: RDNT), today announced a strategic collaboration with CARPL.ai, a leading AI orchestration company that enables radiologists to access, assess, and integrate radiology AI solutions in their workflows. DeepHealth will use CARPL.ai’s technology to develop an AI control system that can be commercialized and will be designed to monitor and optimize imaging AI performance for improved clinical outcomes, operational efficiency, and accelerated adoption of AI in radiology. AI monitoring is crucial to ensure reliable, accurate, and unbiased performance.

The two companies will collaborate on a new closed-loop AI feedback system that will continually monitor AI model accuracy and relevance in clinical settings. The system will automate the measurement and monitoring of performance and safety metrics such as specificity, sensitivity, data- and model drift.

Establishing a robust AI infrastructure with monitoring tools is key for safe, effective, and scalable AI adoption in radiology. While the current landscape is marked by an overwhelming array of AI-enabled point solutions, the future involves running multiple AI models, even for a single use case. DeepHealth’s partnership with CARPL.ai addresses this very need by creating a unique environment to dynamically run a combination of models and monitor performance and then continuously optimize the best models for specific tasks,” said Sham Sokka, PhD, Chief Operating and Technology Officer, DeepHealth.

The partnership will also combine CARPL.ai’s AI marketplace and orchestration platform, which offers a simplified process for selecting, implementing, and monitoring third-party FDA-cleared AI models, with DeepHealth’s cloud-native operating system, DeepHealth OS, which unifies data across the clinical and operational workflows. These platforms will be integrated and extended to monitor real-world workflows on an ongoing basis. The aim is to enable radiologists to access performant and safe AI interpretation tools deeply integrated in their workflows.