Lunit and SimonMed Imaging, a radiology practice network with more than 175 sites in 11 states, have announced a new collaboration to develop and deploy a large-scale custom foundation model for chest radiographs.
According to the companies, Lunit is creating a foundation model, an AI model trained on multimodal datasets containing millions of images to capture generic domain knowledge. SimonMed will fine-tune Lunit’s chest imaging Foundation Model Services (FMS) platform using its own imaging and reports to create a tailored model that reflects the clinical environment of its imaging-center network.
The resultant AI-enhanced workflow will be designed to improve quality, boost speed, reduce variability in reporting and drive greater standardization.
The FMS platform will be continuously improved through model-performance monitoring feedback and drift alerting capabilities.
The chest radiograph model is the first of several, with additional models for mammography and digital breast tomosynthesis planned for deployment in 2026.
