SamenvattingCostly animal/human drug tests over tens of months can sometimes come to a sudden end
because of drug-induced-liver-injury (DILI) caused by hepatotoxicity, leading to restrictions
or even complete suspensions of drugs from the market. To this end, this research project
aims to develop a high throughput detection technique based on Raman read-out for monitoring
liver cells’ behaviour and lipid content inside a microfluidic system, with the final
goal of predicting drug-induced-liver-injury (DILI) as a result of exposure to developed
drugs/chemical compounds in a shorter period.
In this thesis, after a comprehensive overview of DILI and the challenges it poses,
different options for optical detection and cell culturing are investigated. It is shown that
culturing technologies which can maintain a 3D cell culture are desired. Moreover, Raman
spectroscopy is shown to be the best option for dynamic monitoring of biological/chemical
samples. However, Raman signals are usually weak and suffer low signal to noise ratios.
Thus, De-noising algorithms are introduced to improve Raman’s signal to noise ratio. The
results for self-generated and experimental data show that Kalman and Savitzky-Golay
filtering are the best options and improve the limit of detection two times. Finally, the microenvironment
of the cell-culture inside the microfluidic chip is simulated using COMSOL
Multiphysics. It is shown that the flow rates required for cell culturing will not induce any
excessive stress to the cells. Furthermore, the deposited heat as a result of laser radiation will
dissipate sooner than a temperature rise in the cell culture.
The answers to the questions raised by this thesis will form the next stages of the FWO
project entitled “Towards compact, multimodal spectroscopic devices for the read-out of