Radiology AI Gets a Custom Fit: How Two Platforms Are Democratizing Foundation Models for Imaging Departments

Healthcare providers have long faced a dilemma with radiology AI: adopt generic commercial models that don't match institutional workflows, or invest millions in computational resources and proprietary datasets to build custom solutions. A newly announced collaboration between CARPL.ai and HOPPR aims to eliminate that trade-off, offering radiology departments access to powerful foundation models they can fine-tune with their own local imaging data and deploy directly into existing PACS infrastructure.

The partnership, announced October 7, 2026, positions CARPL's platform as the deployment and lifecycle management layer for HOPPR's radiology-specific vision and language foundation models. According to the companies, this combination allows large radiology groups and academic medical centers to develop custom AI applications—including automated report generation aligned with institutional terminology and clinical requirements—without the traditional overhead of building models from scratch.

Breaking Down Barriers to Custom Radiology AI

The collaboration addresses what both companies characterize as critical market gaps in medical imaging AI adoption. Foundation models, which have transformed natural language processing and computer vision in consumer applications, require massive computational resources and proprietary datasets that most healthcare institutions lack. HOPPR's pre-trained radiology foundation models serve as a starting point, enabling organizations to build custom applications using "a fraction of the data" traditionally required, according to the companies.

For radiology departments, this means the ability to create AI tools that reflect their unique reporting styles, terminology preferences, and clinical workflows rather than adapting to generic commercial products. The models integrate directly into existing PACS systems, avoiding the workflow disruptions that have historically hindered AI adoption in imaging departments.

From Procurement to Post-Market Surveillance

CARPL.ai CEO Vidur Mahajan emphasized the platform's role as a comprehensive AI lifecycle solution, stating that the company helps "some of the world's largest radiology groups and academic medical centers adopt AI." The HOPPR partnership extends CARPL's capabilities beyond AI procurement and deployment into custom model development and fine-tuning.

Dr. Khan Siddiqui, Co-Founder, Chairman and CEO of HOPPR, positioned his company as "quickly becoming the undisputed leader in the foundational AI space for radiology." The collaboration leverages HOPPR's specialized imaging models as infrastructure for institution-specific applications rather than end-user products.

What's Next for Custom Imaging AI

Both companies indicated plans to expand the partnership across a wider variety of imaging modalities beyond the initial focus areas. CARPL's existing post-market surveillance tools will be integrated to enable continuous monitoring of custom model performance—a capability the companies describe as critical for maintaining AI accuracy as imaging protocols, equipment, and patient populations evolve.

The partnership signals a broader market shift from one-size-fits-all commercial AI models toward customizable, institution-specific applications built on shared foundation model infrastructure. For healthcare providers, this approach potentially reduces both the technical barriers and financial investment required to deploy AI that fits their radiology workflows, rather than forcing workflows to adapt to generic commercial products.

Sources

1. HOPPR, CARPL.ai Announce Collaboration — ITN Online