31355 Computational fluid dynamics (CFD)-based pre-treatment planning for liver cancer
Begeleider(s): Tim Bomberna

Richtingen: Master of Science in Biomedical Engineering, International Master of Science in Biomedical Engineering

Probleemstelling:

CURRENT SHORTCOMINGS: SUBOPTIMAL TREATMENTS FOR LIVER CANCER 

Hepatocellular carcinoma (HCC) is the most common form of primary liver cancer, ranking fourth in mortality worldwide. At its intermediate, unresectable stage, HCC can be treated by transarterial chemoembolization (TACE) or radioembolization (TARE). During these therapies, catheters are retrogradely advanced through the femoral artery to the hepatic arteries (Fig. 1, panels 1&2), where embolizing microspheres are locally injected to selectively damage tumor tissue (Fig. 1, panel 3). The overall goal is to steer the damaging microspheres towards the tumor tissue to (i) maximize drug delivery to the tumor, and (ii) limit the amount of toxicity delivered to healthy tissue. However, it is currently very challenging to pre-operatively estimate whether microparticles will flow to the tumor after injection. Therefore, methods are needed to pre-operatively and non-invasively assess the suitability of different injection locations and whether or not the catheter needs to be shifted closer to the tumor. 

Figure 1: The workflow for transarterial therapies (TACE, TARE). A catheter is inserted in the femoral artery and navigated towards the liver. Damaging microspheres are injected as close as possible to the tumor to maximize the target-specificity of the treatment.

RESEARCH GOAL: TOWARDS VIRTUAL PRETREATMENT PLANNING  

Therefore, computational fluid dynamics (CFD) modelling has been applied to simulate blood and microparticle flow in patient-specific hepatic arterial geometries. The longer-term goal is to develop a CFD-based pre-treatment planning framework that can estimate the tumor dose for a certain injection location. The added advantage of such a virtual platform is that a clinician can perform different tests before deciding on the final treatment strategy (known as virtual pretreatment planning). Additionally, these CFD models need to be validated before they are used in the clinic. Therefore, this thesis will focus both on the development and validation of such a CFD model of transarterial drug delivery in the liver. 

 


Doelstelling:

The aim of this Master thesis is to develop and validate a patient-specific computer model of transarterial drug delivery in the liver. 

The workflow of this thesis can be divided in several succinct steps.

       (i) Literature study, in which the student(s) will explore the problem setting and identify relevant research questions.

       (ii) Segmentation and meshing of the arterial network of the liver. Practically, imaging data (e.g. CT, MRI) of the liver vasculature of 1-2 patients will be used to create 3D models for simulation geometries (Figure 2) in Mimics (Materialise, Belgium). The meshes (i.e. subdivision of the geometry in small elements) can be made in Fluent Meshing (Ansys, USA). 

Figure 2: Two examples of patient-specific liver geometries which were extracted from micro-CT images. 

        (iii) CFD modelling. From a modelling point of view, this transport problem can be translated in a dispersion of discrete phase particles (drug carriers) in a continuous fluid phase (blood). This will be done in Fluent (Ansys, USA). Using this model, the number of particles flowing to the tumor can be determined. 

        (iv) Validation. Validation can be done by calculating the dose delivered to the tumor and overlapping numerical results with SPECT-CT images. The radioactive hotspots show where microspheres deposited in the tissue during the (pre-)treatment, which should correspond with particle deposition as predicted by the model (Figure 3).   

Figure 3: Patient-specific validation using SPECT-CT imaging. 

The described thesis subject is broad and is open for input from the student(s). For example, the student(s) can choose to focus on different aspects of the modelling process (i.e. segmentation, validation, wide parameter study, ...) according to his/her/their own interests. The subject is ideal for students interested in computational modelling as a tool to make therapies more patient-specific and improve treatment outcomes (known as in silico medicine).