Author: Aditya P. Apte, Joseph O. Deasy, Sharif Elguindi, Aditi Iyer, Jue Jiang, Eve Marie LoCastro, Jung Hun Oh, Amita Shukla-Dave, Harini Veeraraghavan 👨🔬
Affiliation: Department of Medical Physics, Memorial Sloan Kettering Cancer Center, Memorial Sloan Kettering Cancer Center 🌍
Purpose: We present port of popular Computational Environment for Radiological Research software platform to Python programming language to cater to cloud-based analyses.
Methods: The components of pyCERR can be broadly categorized as (1) Data container (2) Visualization and (3) Analysis modules. It builds on commonly used scientific libraries Numpy, scipy, pandas and image processing libraries such as, pydicom, SimpleITK, scikit-image. pyCERR uses Napari for Visualization to allow access to graphical user interface object handles and their programmatic manipulation in Python. The plan container class planC houses data classes containing attributes for various data types. These include (a) Scan for imaging data such as CT, MR, PET, (b) Structure for image segmentation, (c) Dose for RT dose, (d) Beams for RT Plans as well as derived data types such as deformable image registration.
Analysis modules facilitate cloud-compatible AI-based workflows for image segmentation, radiomics, DCE MRI analysis, radiotherapy dose-volume histogram-based features, and normal tissue complication and tumor control models for radiotherapy. Image processing utilities are provided to help train and infer convolutional neural network-based models for image segmentation, registration and transformation. The framework enables users to apply AI models to their images on a cloud platform such as Cancer Genomics Cloud (CGC) and retrieve and review results on their local machine without requiring local installation of specialized software or GPU hardware. The deployed AI-based workflows are accessible using API provided by CGC in a variety of programming languages.
Results: pyCERR is distributed as an open-source, GNU-copyright software on GitHub along with Notebooks demonstrating various applications for AI segmentation, visualization, radiomics and radiotherapy feature calculation.
Conclusion: pyCERR facilitates end-to-end radiological image analysis and reproducible research, including pulling data from cloud sources, training or inferring from an AI model, utilities for data management, visualization, and simplified access to image metadata.