Learning neuroimaging models from health system-scale data

2025-09-23

AI Summary

Escalating clinical demand for MRI exacerbates healthcare resource disparities, particularly in primary-care and rural settings. To address this, we propose Prima—the first vision-language model (VLM) capable of processing real-world clinical MRI inputs. Trained on 220,000 multicenter, multi-device, and multi-population MRI scans paired with radiology reports, Prima introduces a hierarchical visual architecture enabling robust, disease- and scanner-agnostic feature extraction. It further integrates interpretable diagnostic reasoning, intelligent worklist prioritization, and evidence-informed referral recommendations. Evaluated on an independent test set of 30,000 cases, Prima achieves a mean AUC of 92.0% across 52 neurological disorders—significantly outperforming existing methods—while demonstrating strong generalizability and algorithmic fairness across demographic and technical subgroups. Prima thus delivers a deployable AI solution to alleviate bottlenecks in radiology service capacity.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Hardware-aware ML
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Neuroimaging is a ubiquitous tool for evaluating patients with neurological diseases. The global demand for magnetic resonance imaging (MRI) studies has risen steadily, placing significant strain on health systems, prolonging turnaround times, and intensifying physician burnout cite{Chen2017-bt, Rula2024-qp-1}. These challenges disproportionately impact patients in low-resource and rural settings. Here, we utilized a large academic health system as a data engine to develop Prima, the first vision language model (VLM) serving as an AI foundation for neuroimaging that supports real-world, clinical MRI studies as input. Trained on over 220,000 MRI studies, Prima uses a hierarchical vision architecture that provides general and transferable MRI features. Prima was tested in a 1-year health system-wide study that included 30K MRI studies. Across 52 radiologic diagnoses from the major neurologic disorders, including neoplastic, inflammatory, infectious, and developmental lesions, Prima achieved a mean diagnostic area under the ROC curve of 92.0, outperforming other state-of-the-art general and medical AI models. Prima offers explainable differential diagnoses, worklist priority for radiologists, and clinical referral recommendations across diverse patient demographics and MRI systems. Prima demonstrates algorithmic fairness across sensitive groups and can help mitigate health system biases, such as prolonged turnaround times for low-resource populations. These findings highlight the transformative potential of health system-scale VLMs and Prima's role in advancing AI-driven healthcare.

Problem

• Addressing rising MRI demand and physician burnout in neurological disease evaluation
• Improving diagnostic accuracy for diverse neurological disorders using neuroimaging data
• Mitigating healthcare system biases and prolonged turnaround times for underserved populations

Innovation

• Vision language model trained on 220,000 MRI studies
• Hierarchical vision architecture for general MRI features
• Provides explainable diagnoses and prioritizes radiologist worklist