Philips and NVIDIA Collaborate to Enhance MRI Predictive Imaging

Philips, in partnership with NVIDIA, is developing technology to advance MRI imaging through predictive previews powered by accelerated computing and AI models. The initiative aims to streamline scan planning and promote more autonomous imaging workflows.

MRI technologists face numerous challenges before a scan, such as selecting protocols, positioning equipment, and anticipating patient-specific issues like motion or artifacts. Philips seeks to simplify this process by using AI to create an image preview before data acquisition. This preview can validate scan plans and guide setup decisions, moving MRI towards more consistent, high-quality exams.

The collaboration with NVIDIA involves developing a domain-specific MRI Foundation Model, trained on extensive MRI data covering various anatomies, protocols, and contrast settings. This model, combined with NVIDIA's NV‑Generate‑MR, NV‑Segment, and NV‑Reason models, seeks to integrate image generation, segmentation, and interpretation into a single workflow. By learning the complexities of MRI, such as anatomical changes and artifact recognition, this model can produce generalized representations for specific applications.

Sathish Kumar Balakrishnan, Head of Global Research and Development for MRI at Philips, stated, “By combining our MR foundation model with NVIDIA’s AI solutions, we see an opportunity to make scan setup feel more guided and consistent for technologists.”

Predictive image previews enable operators to assess protocol and positioning choices before scanning begins, potentially reducing rescans and improving overall image quality. Such tools can simplify decision-making and promote uniform results across different operators and facilities, enhancing operator confidence and patient care.

The demand for advanced diagnostic imaging continues to rise, amid staffing shortages and increasing patient numbers. By designing MRI systems with more autonomous features, Philips aims to simplify this complexity, thus delivering consistent high-quality imaging exams across various settings. This initiative promises to improve patient diagnostics and reduce the burdens on healthcare systems, ensuring timely and efficient patient care.