HOPPR Introduces Beta FS MSK Whole-Body Narrative Model for Radiology Reporting

HOPPR has announced the beta release of its FS MSK Whole-Body Narrative Model, a vision-language foundation model aimed at generating draft findings text for musculoskeletal (MSK) radiographs. This model can process findings from up to three views per study, providing a software component that developers can use to build radiology reporting and image-based workflow applications.

The FS MSK Model has been trained across various regions of the musculoskeletal system, including the cervical spine, thorax, chest wall, upper extremities, pelvis, and lower extremities. The training utilized about 674,000 proprietary outpatient MSK studies, ensuring a balanced dataset across different anatomical parts. The focus of the model is on identifying abnormalities rather than normal anatomical features, with performance evaluated across 26 MSK finding categories such as fractures, degenerative joint changes, alignment, soft tissue, bone quality, lesions, and hardware evaluation.

Khan Siddiqui, MD, co-founder and CEO of HOPPR, stated, "Organizations shouldn't need a separate AI model for every bone and joint. Musculoskeletal imaging spans the whole body, and a single model trained across it gives developers a far stronger foundation to build on." He emphasized that the FS MSK Model provides a comprehensive solution for draft findings text generation, which can be customized to fit individual organizations' needs through HOPPR Forward Deployed Services (FDS).

The FDS team, comprising machine learning scientists, software engineers, data scientists, and clinical experts, collaborates closely with organizations to customize the model with their own data, design suitable labeling workflows, integrate the model into existing systems, and support validation testing in specific environments. HOPPR's secure infrastructure is HITRUST e1 certified and SOC 2 Type II attested, ensuring compliance with quality management standards.

The model's name, "FS," is a tribute to Dr. Florence Ada Stoney, widely recognized as the first woman radiologist in the United Kingdom. Roger Boodoo, MD, Medical Director of AI at HOPPR, remarked on the significance of naming the model in her honor, acknowledging the pioneers in radiology during its early development.

It is important to note that the FS MSK Whole-Body Narrative Model is a software component and is not classified as a medical device. It is intended for evaluation and feedback purposes only and is not cleared or approved by the U.S. FDA for diagnostic use. Organizations using this model are responsible for validating its performance and securing any necessary regulatory approvals for commercial use.