BME Research Day 2018

The BME Research Day is the official annual event of the Biomedical Engineering department of TU/e, where students, PhDs, PostDocs, and Researchers come together, showcase their research with lectures and a dedicated poster session, and network in a stimulating environment.

The organizing committee of the BME Research Day of this year consisted of Tom de Greef, Associate Professor of Synthetic Biology, Patricia Dankers, Full Professor of Biomedical Materials, Menno Prins, Full Professor at the Molecular Biosensors for Medical Diagnostics group, and Rina van Dijck, secretary of the Medical Image Analysis group. Much of the attention of this year’s edition was devoted, amongst others, to cancer screening, diagnostic and treatment, and to the Mechanobiology of load bearing tissues such as tendon, ligaments and cardiovascular tissues.

Microfluidics and Predictive modelling for Cancer screening and treatment: towards personalized medicine

Kicking off the morning session was the lecture of Federica Aduati on “Predictive logic models of signalling pathways from microfluidics-based drug screening of tumour biopsies”. Aduati, who was recently appointed Assistant Professor in the Computational Biology group, has a background in Bioengineering, with particular focus on precision medicine and signaling pathways in cancer therapy from both a computational and an experimental perspective.

Aduati: “Today, patients diagnosed with cancer usually receive the same treatment as others who have same type and stage of cancer. However, scientists now understand that patients’ tumors have genetic changes that cause cancer to grow and spread”. In other words, changes that occur in one person’s cancer may not occur in others who have the same type of cancer.

“So why not to measure the effect of specific drug treatments directly on patients’ sample?” says Aduati. While this sounds as the natural solution to the problem, the road to success is still full of obstacles. “This approach has inherent limitations”, explains Aduati, “such as the shortage of samples, the limited set of conditions that can be tested and, ultimately, the need for ex vivo culturing, which is quite labor intensive”. According to Aduati, phenotypic screening of biopsies via microfluidics could be a promising solution in that respect. “Cells from patients can be encapsulated with drugs and fluorescent markers and monitored over time”, explains Aduati. To date, the model developed by Aduati in collaboration with the European Molecular Biology laboratory (EMBL), the EMBL-European Bioinformatics Institute (EBI) of Aachen, the University of Aachen and the University of Maastricht, was validated successfully using 2 pancreatic cell lines and on mice models in vivo, and successfully applied for the screening of 4 pancreatic tumor biopsies from patients.

Learning from small data in medical image analysis: machine learning algorithms for cancer screening and diagnostic

Following the talk of Aduati, was the lecture of Veronika Cheplygina, Assistant professor at the Medical Image Analysis group. The research of Cheplygina focuses on machine learning in medical image analysis, and on scenarios where not enough annotated data are available. “Imagine having a big scan of one entire organ” - explains Cheplygina - “which contains different tissues and a mixed composition of healthy and abnormal areas. It would be beneficial to keep those areas separate, and have two separate algorithms learning from these two different scenarios.” One solution might rely on multiple instance learning, or, in other words, the possibility for the algorithms to learn with global information extracted from ambiguous data. Alternatively, transfer learning could be used, which focuses on learning and storing knowledge while solving problems, and applying that knowledge to different, yet related, problems. Ultimately, a combination of human insights and machine learning techniques could be exploited, in the so-called crowdsourcing approach. “This is exactly what Facebook does, for example, when learning how to recognize faces whenever you tag someone in a picture”, explains Cheplygina. The algorithms developed by Cheplygina can learn from examples and make predictions about novel data. Specifically, by learning from medical scans with annotated abnormalities, these algorithms might detect abnormalities in previously unseen patients, opening up for more effective ways to screen and diagnose cancer.

Predicting drug response in breast and prostate cancer: challenges and opportunities

Keynote speaker for this year edition of the BME event was Dr. Wilbert Zwart from The Netherlands Cancer Institute (NKI).  The research interests of Dr. Zwart lie at the crossroad between genomics, pathology, cell biology and molecular pathology. “Combining specialities from different research fields is key for innovative translational science in cancer research” says Dr. Zwart. “The hormone-related mechanisms in transcriptional regulation of breast cancer and prostate cancer are tightly linked”, explains Dr. Zwart, which calls for a scientific approach to the problem based on the synergy between different research lines. The central theme of the research projects of Dr. Zwart’s team revolves around the development of personalized cancer treatment, the identification of novel biomarkers and the possibility to unravel the underlying biological mechanisms occurring in cancer.

Tendon and Ligament Injuries: Replacement medicine or Regenerative medicine?

Active, sporty people have 1% risk of anterior cruciate ligament (AIL) rupture. “During surgery” – explains Jasper Foolen, Assistant Professor at the Orthopaedic Biomechanics group – “torn ACL is generally replaced by a substitute graft made of tendon. This procedure might have a lot of complications, such the graft’s rupture over time, knee instability, and joint degeneration. It is estimated indeed that only for 50 % of the patients undergoing surgery can go back to a normal and active lifestyle.” To understand the underlying causes, Foolen and his co-workers went back to how the surgery is performed, particularly focusing on the tissue remodeling process occurring in vivo. “When surgeons remove the original tendons, they scrape off the muscle, leaving behind a disorganized mesh of fibers, which cannot absorb and distribute the mechanical loads as a healthy and fully functional tendon would do”. To understand how this process occurs, Foolen and his colleagues performed mechanical testing on explanted tendons, confirming a substantial decay of the mechanical properties after tissue deprivation.

Beside AIL, the research interests of Foolen encompass also the study of tendinopathy and the remodeling potential of tendon-derived cells.  “Tendinopathy associates with a shift from healthy tissue with aligned extracellular matrix (ECM) towards a diseased tissue with a disorganized ECM and randomly distributed cells with scar-like features” explains Foolen. A fundamental clinical dilemma with this scarring process is whether treatment strategies should focus on healing the scarred tissue, or strengthen instead the remaining healthy tissue. To answer this question, Foolen and colleagues developed a next generation tissue platform that mimics cellular responses during tendon scarring. “We demonstrated that cells residing in an unscarred anisotropic tissue have superior remodeling capacity, when compared to their scarred isotropic counterparts”, says Foolen, which would lend to support strategies focusing on strengthening the remaining healthy tissue, rather than regenerating scarred tissues.

“The BME research day as a valuable experience to present in front of a wide audience of faculty members and students”

This edition of the BME Research Day was also the perfect occasion for MSc, PhD students, PDEng trainees and postdocs of the Department to present their research to a broad scientific audience via poster sessions and  pitches. This year, the poster price was won by the master student Eva van Aalen. Johanna Melke, postdoc at the Orthopaedic Biomechanics group, Emilien Dubuc, postdoc in Computational Biology, and Nicole van Gestel, PhD candidate at the Orthopaedic Biomechanics group shared the second place.

Melke: “The BME research day gave me valuable presentation experience and it was a great opportunity to present my work to a wide audience of interested faculty members and students. Showcasing a lot of different disciplines, it was definitely a great way to come together, discuss research, and learn something new.”

Next year edition of the BME Research Day will take place in April 2019. Stay tuned for more information on the program and the registration process.