Radiomics Cancer Detection Platform

I want a production-ready application that lets clinicians upload CT, MRI, and PET scans, automatically extract radiomics features, run an AI/ML classifier for early cancer detection, and immediately generate an interpretable report they can save or share.

Core workflow
• Secure image upload (DICOM or NIfTI) with basic anonymisation
• Automated or manual region-of-interest selection, followed by feature extraction (PyRadiomics or equivalent)
• Model inference using a configurable TensorFlow/PyTorch pipeline and clear probability scores
• Interactive visualisation of key features and predictive explanations
• One-click export of findings to PDF/JSON and storage in a local or cloud database

Acceptance criteria
1. End-to-end processing of CT, MRI, and PET studies in under five minutes per case on a mid-tier GPU.
2. Feature extraction and classification accuracy reproducible from a provided test set and Jupyter notebook.
3. Modular, well-documented codebase with environment files and unit tests.
4. Installer (or Docker image) plus user manual so a radiologist can be fully operational in one hour.

Please attach a detailed project proposal outlining architecture, preferred libraries, timeline with milestones, and any prior experience delivering medical imaging or radiomics solutions.