HOPPR taps NVIDIA for medical AI
HOPPR™ AI Foundry expanded its medical imaging toolbox by integrating NVIDIA’s NV‑Reason and NV‑Generate foundation models, making high‑throughput image interpretation and model development more accessible to pathology teams. That move accelerates plug‑and‑play AI for cell detection, triage, and automated QA across digital cytology projects. (prnewswire.com)
HOPPR announced the addition of NVIDIA’s NV‑Reason and NV‑Generate to the HOPPR™ AI Foundry on March 17, 2026 during NVIDIA GTC 2026 (prnewswire.com ). NV‑Reason is shipped as a multimodal “reasoning” vision‑language model that produces structured, chain‑of‑thought style explanations alongside image outputs, with an existing NV‑Reason‑CXR‑3B model focused on chest X‑ray interpretation. (developer.nvidia.com huggingface.co ). NV‑Generate is a family of 3D latent diffusion models (examples: NV‑Generate‑CT, NV‑Generate‑MR, NV‑Generate‑MR‑Brain) designed to produce high‑fidelity synthetic DICOM volumes and paired segmentation masks for data augmentation and rare‑case synthesis. (huggingface.co huggingface.co github.com ). The Foundry deployment specifies NVIDIA A100 and H100 GPUs for large‑scale training and inference, and supports performance tooling including NVIDIA TensorRT optimization, NVIDIA Triton Inference Server, and MONAI‑based transforms for imaging pipelines. (prnewswire.com ). HOPPR describes the Foundry as operating under a Quality Management System aligned to ISO 13485, IEC 62304, ISO/IEC 42001, and ISO 14971, and the platform includes a curated private dataset totaling more than 19 million studies sourced from 18 partner imaging sites. (medicaldesignandoutsourcing.com ). Model customization and validation workflows on Foundry allow running inference on NVIDIA open models and fine‑tuning them with hands‑on support from HOPPR Forward Deployed Services (FDS) inside a HIPAA‑compliant development environment. (prnewswire.com ). Press coverage from imaging trade outlets highlights NV‑Reason’s structured analytical outputs and NV‑Generate’s role in creating synthetic datasets where real‑world data are limited, framing the HOPPR–NVIDIA integration as providing explainability and synthetic‑data tooling for traceable model validation and QA workflows. (auntminnie.com huggingface.co ).
