Medical physics · 2026
Artificial intelligence apps for medical image analysis using pyCERR and cancer genomics cloud.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Apte Aditya P, LoCastro Eve, Iyer Aditi, Elguindi Sharif, Jiang Jue, Thor Maria, Oh Jung Hun, Veeraraghavan Harini, Shukla-Dave Amita, Deasy Joseph O
Аннотация
PURPOSE: To provide a software resource that enables medical researchers and practitioners to develop and apply reproducible, out-of-the-box, AI-based analysis workflows for radiological and radiotherapy data.
DEVELOPMENT AND VALIDATION METHODS: The framework comprises two main components: the Python-native computational environment for radiological research (pyCERR) for radiological image processing, and the cancer genomics cloud (CGC), a National Cancer Institute-funded resource that supports cloud-hosted collaborative projects, with tools for managing collaborators and compute-hardware from providers such as Amazon Web Services and Google Cloud Platform. pyCERR is open-source (GNU GPL) and distributed at https://github.com/cerr/pycerr; it provides user-friendly data structures, simplified access to metadata across radiological imaging modalities, and tools for data transformation, AI-based image segmentation, radiomics, dynamic contrast-enhanced MRI analysis, radiotherapy dose-volume histogram (DVH) feature extraction, and normal-tissue complication and tumor-control models. Radiomics features comply with the Image biomarker standardization initiative (IBSI) standards and are validated against published reference values. Unit tests covering data input/output, radiomics, DVH features, and visualization run on every code update. Workflows are containerized for reproducible, identical execution on the CGC and on local hardware. The CGC enables AI inference from radiological images to be combined with other data types such as genomics and proteomics. Our public CGC project (https://cgc.sbgenomics.com/u/sevenbridges/pycerr-radiological-radiotherapy-imageanalysis), accessible with free registration, currently includes apps for organ-at-risk (OAR) and lung-nodule segmentation based on published models trained in-house and clinically validated. These apps can be run interactively from a web browser or programmatically via a REST API with client support for multiple programming languages. Notebooks demonstrating the CGC API and pyCERR capabilities are provided at https://github.com/cerr/pycerr-notebooks.
POTENTIAL APPLICATIONS: The framework facilitates end-to-end radiological image analysis and reproducible research, including data access from public and private repositories, training and inference with AI models, and utilities for data management, visualization, and simplified access to image metadata.
DEVELOPMENT AND VALIDATION METHODS: The framework comprises two main components: the Python-native computational environment for radiological research (pyCERR) for radiological image processing, and the cancer genomics cloud (CGC), a National Cancer Institute-funded resource that supports cloud-hosted collaborative projects, with tools for managing collaborators and compute-hardware from providers such as Amazon Web Services and Google Cloud Platform. pyCERR is open-source (GNU GPL) and distributed at https://github.com/cerr/pycerr; it provides user-friendly data structures, simplified access to metadata across radiological imaging modalities, and tools for data transformation, AI-based image segmentation, radiomics, dynamic contrast-enhanced MRI analysis, radiotherapy dose-volume histogram (DVH) feature extraction, and normal-tissue complication and tumor-control models. Radiomics features comply with the Image biomarker standardization initiative (IBSI) standards and are validated against published reference values. Unit tests covering data input/output, radiomics, DVH features, and visualization run on every code update. Workflows are containerized for reproducible, identical execution on the CGC and on local hardware. The CGC enables AI inference from radiological images to be combined with other data types such as genomics and proteomics. Our public CGC project (https://cgc.sbgenomics.com/u/sevenbridges/pycerr-radiological-radiotherapy-imageanalysis), accessible with free registration, currently includes apps for organ-at-risk (OAR) and lung-nodule segmentation based on published models trained in-house and clinically validated. These apps can be run interactively from a web browser or programmatically via a REST API with client support for multiple programming languages. Notebooks demonstrating the CGC API and pyCERR capabilities are provided at https://github.com/cerr/pycerr-notebooks.
POTENTIAL APPLICATIONS: The framework facilitates end-to-end radiological image analysis and reproducible research, including data access from public and private repositories, training and inference with AI models, and utilities for data management, visualization, and simplified access to image metadata.