Open projects
We have several open research projects for MSc students who want to do an internship or graduation project. We keep these projects in a portfolio document (button below). In general, the projects focus on medical image processing.
BIGR MSc Project Portfolio
Research lines projects
Artificial Intelligence for Integrated Diagnostics (AIID)
-
Diagnosing liver lesions using multiparametric MRI requires detecting complex, often subtle features - such as washout, fat content, capsular enhancement, and iron deposition - However, these features can be subtle and challenging to detect, often requiring the expertise of experienced radiologists, which limits the scalability of supervised AI approaches. With the rise of vision-language models (VLMs) like GPT-4V and DeepSeek-VL, a key question is whether such models can reliably identify and describe these features across diverse MRI sequences. This project explores the diagnostic potential of VLMs as assistive tools, with a particular focus on benchmarking their ability to see and reason over relevant imaging features in comparison to expert radiologist interpretations and pathology- confirmed diagnoses. The primary goal is to systematically evaluate how well these models extract and describe clinically meaningful features from annotated MRI data, and to what extent their predictions aligns with real-world diagnostic outcomes. Depending on results and interest, the project may also explore fine-tuning existing models to improve domain-specific performance, or - if performance proves reliable - developing a lightweight graphical interface to make these models more accessible and interpretable in a clinical context. This research is part of the Liver Artificial Intelligene - Consortium, a collaborative initiative focused on advancing AI models in liver imaging, bringing together interdisciplinary expertise to unlock new diagnostic possibilities.
-
Desmoid tumors, also known as aggressive fibromatosis, are rare soft tissue neoplasms that exhibit locally invasive growth but lack metastatic potential. While not typically life-threatening, their infiltrative nature can lead to significant morbidity depending on their location. A large proportion of sporadic desmoid tumors harbor activating mutations in the CTNNB1 gene, which encodes the β-catenin protein. Among these, the S45F mutation is of particular clinical interest as it is associated with increased tumor aggressiveness, higher recurrence rates, and reduced response to conservative therapies.Routine detection of the S45F mutation currently relies on molecular assays such as PCR or sequencing, which, although accurate, are time-consuming, costly, and not always available in low-resource settings. In contrast, hematoxylin and eosin (H&E) stained histopathology slides are inexpensive and routinely available in all diagnostic workflows. Recent advances in computational pathology have shown that deep learning models can infer underlying molecular alterations from subtle morphologic features in H&E slides. This MSc project aims to develop and evaluate a deep learning model that predicts the presence of the S45F mutation in desmoid tumors directly from digitized H&E-stained slides. Such a model could serve as a cost-effective screening tool to complement molecular testing, reduce unnecessary assays, and enhance decision-making in clinical practice. For more information contact: m.starmans@erasmusmc.nl or k.prathaban@erasmusmc.nl.
-
AI models in medical imaging are usually developed from scratch using disease-specific datasets. While this works for common diseases, it poses a major bottleneck for rare cancers like soft-tissue tumors (STTs), where labelled data is scarce. Unlike humans, who can learn new concepts from just a few examples by drawing on prior experience, current AI systems require large amounts of task-specific data due to how they are trained. This project addresses that limitation by combining meta-learning and automated machine learning (AutoML) to enable knowledge transfer across clinical and reducing reliance on large datasets.
-
Soft tissue tumors comprise a rare but diverse and diagnostically challenging group of neoplasms with varying histological subtypes, biological behavior, and clinical outcomes. Accurate and scalable phenotyping from histopathology slides is essential for improved diagnosis, grading, and treatment planning. Traditional pathology relies heavily on expert interpretation of H&E-stained tissue sections, but the increasing availability of digital pathology and large-scale annotated datasets presents a unique opportunity to apply deep learning for tumor characterization. This MSc project contributes to the development of a large-scale AI model for the phenotyping of soft tissue tumors directly from whole slide images (WSIs). The project is part of a broader effort to build a robust and generalizable model that can later be fine-tuned for downstream tasks such as prognosis prediction, molecular alteration inference, and treatment response modeling. Students will have the opportunity to explore various deep learning approaches, including supervised and self-supervised representation learning, weak supervision, and multi-task learning. The project can also focus on a specific tumor type (e.g., lipomas/liposarcomas or GISTs), depending on interest and data availability. It offers hands-on experience with high-resolution pathology data, modern AI toolkits (e.g., PyTorch, MONAI), and opportunities for contribution to a high-impact clinical research direction. For more information contact: m.starmans@erasmusmc.nl or k.prathaban@erasmusmc.nl.
Biomedical Imaging Research Infrastructure
If you have your own ideas for a medical imaging analysis MSc project, feel free to contact dr.
Stefan Klein.