Colorectal cancer research group

Our research group applies an interdisciplinary approach to colorectal cancer. We investigate the relevance of tumor budding and the tumor microenvironment in different clinical scenarios, and apply computational methods (e.g. deep learning, graphs) to discover new features, train models and eventually validate and translate findings in routine setting.
Current research projects Lugli Zlobec
Digital pathology and machine learning

Group Lugli Zlobec Our group uses digital pathology and machine learning to investigate clinically relevant scenarios in colorectal cancer and the tumor budding/microenvironment using histopathological images and ngTMAⓇ. We apply deep learning to identify genotype-phenotype correlations (post-doc work by Dr. HG. Nguyen) and develop algorithms for future validation and use in diagnostic routine (PhD project A. Khan). Together with the lab of Prof. J.-P- Thiran (EPFL), we develop computational methods for prognosis classification (PhD project C. Abbet) and with Prof. A. Fischer (UniFribourg/HESO) apply graph-representations and geometric deep learning to understand the tumor microenvironment (PhD project L. Studer). Our numerous Master students in Bioinformatics (MSc) and Medicine (MMed) work on interdisciplinary projects, all on the basis of histological images.
Example of p53 in an endometrial cancer within an ngTMA experiment
The epigentic landscape of CDX2

Group Lugli Zlobec Up to 20% of colorectal cancers have decreased expression or complete loss of CDX2 protein. CDX2 loss is associated with microsatellite instability, high-level CpG island methylation and BRAF mutation, features consistent with the serrated pathway. Our functional studies show that hypermethylation of the CDX2 promoter is a major reason for this loss and can be recovered with DNMTi treatment. Together with the group of Prof. M. Tschan, we investigate genetic and epigenetic modifications of CDX2 using CRISPR/Cas9 technology.
A) Positive expression of CDX2 protein in colorectal cancer. B) Loss of CDX2 has a significant negative impact on patient outcome (Kaplan-Meier curve, log-rank test)
Tumor Budding in Colorectal Cancer

Group Lugli Zlobec Tumor buds are linked to aggressive tumor behaviour and poor prognosis. One of our objectives is to standardize the assessment and reporting of tumor budding in routine diagnostics. This was the main incentive for our group to initiate and organize the first International Tumor Budding Consensus Conference (ITBCC) which took place in 2016. The proceedings of the ITBCC were published in the journal ‘Modern Pathology’ in 2017 and have since been integrated in reporting protocols of major societies such as the College of American Pathologists. One of our ongoing projects is to validate the ITBCC recommendations on large colorectal cancer cohorts, also in collaboration with the International Budding Consortium (IBC). Other aspects of our research are focused on further characterization of tumor buds in primary colorectal cancer and colorectal liver metastases within the tumor microenvironment. Our ultimate goal is to find targets that can be used against tumor buds in a therapeutic setting.
Tumor buds in colorectal cancer (H&E staining, 40x)