Medical Imaging Lab
MONAI + TotalSegmentator + Pillar-0 via API. Research-grade medical image analysis without managing CUDA clusters.
What's inside
TotalSegmentator + MONAI Segmentation
Segment 100+ anatomical structures in CT, MRI, and X-ray. TotalSegmentator is the current state-of-the-art open-source segmentation framework. MONAI is the PyTorch of medical imaging.
Pillar-0 Classification
UC Berkeley/UCSF's open-source Pillar-0 model. Outperforms Google MedGemma and Microsoft MI2 across 350+ findings. Radiology, pathology, and dermatology classification in one endpoint.
Anomaly Detection
Unsupervised anomaly detection in medical images. Flag regions of interest for downstream review. Built on MONAI + nnU-Net.
Draft Radiology Reports
Generate research-grade draft radiology reports for training and data annotation. Never for clinical decision making.
FAQ
IS THIS FDA CLEARED?›
No. SciRouter's Medical Imaging Lab is for research, education, and data annotation only. It is NOT a diagnostic tool, NOT FDA-cleared, and must NOT be used for clinical decision making. Every response includes a disclaimer.
What models are supported?›
TotalSegmentator v2 for anatomy segmentation, Pillar-0 for classification, MONAI for anomaly detection, and nnU-Net-derived models for task-specific segmentation. All open-source, research-grade.
What about patient privacy?›
You control the data. SciRouter processes images in-memory and returns results. For HIPAA-compliant workflows, self-host the inference stack. We don't log image content, but we do log request metadata (endpoint, latency, credits).
Can I use this for my research paper?›
Yes — please cite SciRouter + the underlying model authors (MONAI, TotalSegmentator, Pillar-0). The tools are open-source; SciRouter provides the API layer.
What's the TAM?›
FDA has approved 873+ AI radiology algorithms. The AI medical imaging market is projected at $7B+ by 2033 (38% CAGR). Research institutions, MedTech startups, and academic hospitals are the primary audience.