Related Experiment Video
Updated: May 13, 2026
10:27
Testing Targeted Therapies in Cancer using Structural DNA Alteration Analysis and Patient-Derived Xenografts
Published on: July 25, 2020
TumorTwin: a Python framework for patient-specific digital twins in oncology
Michael G Kapteyn1, Anirban Chaudhuri2, Ernesto A B F Lima2,3
1Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, TX, USA. michael.kapteyn@austin.utexas.edu.
BMC Medical Informatics and Decision Making
|May 12, 2026
Summary
TumorTwin is a new software framework for creating patient-specific cancer digital twins. It enables dynamic recalibration and supports clinical decisions for optimal cancer treatment strategies.
Area of Science:
- Computational oncology
- Digital twin technology
- Medical imaging analysis
Background:
- Computational oncology advances enable patient-specific tumor growth and treatment response prediction.
- Digital twin frameworks integrate physical tumor data for dynamic recalibration and decision support.
- Existing digital twin frameworks often require bespoke implementations for specific diseases and models.
Purpose of the Study:
- To introduce TumorTwin, a modular and differentiable software framework for patient-specific cancer digital twins.
- To provide a flexible and reusable infrastructure for computational oncology research.
- To facilitate the development and testing of image-guided oncology digital twins.
Main Methods:
- Developed a modular and differentiable Python package named TumorTwin.
- Created an adaptable patient-data structure for diverse disease sites.
- Implemented a modular architecture for composing data, model, solver, and optimization components.
- Enabled CPU or GPU parallelized forward model solves and gradient computations.
Main Results:
- Demonstrated TumorTwin's functionality using an in silico dataset of high-grade glioma growth and radiation therapy response.
- Showcased the framework's ability to initialize, update, and leverage patient-specific digital twins.
- Highlighted the public availability of TumorTwin with comprehensive documentation and tutorials.
Conclusions:
- The TumorTwin framework accelerates the prototyping and testing of image-guided oncology digital twins.
- Enables systematic investigation of various models, algorithms, disease sites, and treatment decisions.
- Leverages robust numerical and computational infrastructure for advanced cancer research.
