Predicting CTNNB1 S45F Mutation in Desmoid Tumors from H&E-Stained Histopathology Images
AI for desmoid CTNNB1 mutation prediction on digital pathology
Background
Desmoid tumors are rare soft tissue neoplasms that exhibit local invasiveness without metastatic potential. A significant subset of sporadic desmoid tumors harbors activating mutations in the CTNNB1 gene, particularly the S45F mutation, which is associated with higher recurrence risk and poorer response to conservative therapies. Detecting this mutation currently requires molecular assays, which are costly and not always available in resource-limited settings. This project explores the potential of using deep learning to predict S45F mutation status directly from histology images, enabling a scalable screening alternative.
Aim
he aim of this project is to develop and validate a deep learning model that predicts the presence of the CTNNB1 S45F mutation in desmoid tumors from H&E-stained whole slide images. The model will be trained using annotated data with known mutation status and will explore image-based features that correlate with underlying molecular changes. The project will involve data preprocessing, model development (e.g., using convolutional neural networks or vision transformers), and evaluation of predictive performance. This work has the potential to improve personalized decision-making in desmoid tumor management.
Related research:
Supervisors
- Karthik Prathaban
- Farhan Akram
- Martijn Starmans
Feel free to mail me if you are interested in this project or want more information!
Martijn P. A. Starmans
Assistant Professor & PI Artificial Intelligence for Integrated Diagnostics (AIID) focused on Medical Imaging in Oncology
My research interests include radiomics, pathomics, multimodal machine-learning, AutoML, and meta-learning.