Targeted drug delivery for liver cancer: a patient-specific computational model of the particle transport during radio-embolisation

Abstract

Hepatocellular carcinoma is the most common liver malignancy and it is predicted to grow to 22 million cases over the next two decades. Depending on the stage of the tumour, the therapy is chosen based on the Barcelona clinic liver cancer (BCLC) criteria. Transarterial therapies are among the possible treatments and they can be divided into transarterial embolization (TAE), transarterial chemoembolization (TACE) and radioembolization (RE). These therapies take advantage of the fact that the tumours are mostly fed by hepatic arteries instead of by hepatic veins as healthy tissue does. In this context, Computational Fluid Particle Dynamics (CF-PD) have been proven to be a powerful tool to evaluate how to maximize the targeting to the tumours as well as to asses which parameters play the most significant role in the final distribution of the drugs. Also, this technique can be used to generate particle release maps (PRMs) that link the injection points with the path that the particles will follow. Thus, they can be used to decide the injection point to target the tumours.
In this work, the possibility of targeting the tumours in a patient-specific geometry by means of CF-PD has been evaluated as well as the parameters that play the most significant role in the final distribution of the drugs. For this purpose, the geometry of the patient was extracted from Magnetic Resonance Images (MRI) by means of segmentation and computer simulations were run in it. For this patient-specific geometry, it was proven that assessing which treatment would provide the easiest target to the tumours is more than feasible by means of CF-PD. Also, that the injection location plays a significant role in terms of particle distribution and the particle density and particle diameter in terms of distal penetration. Even if the particle distribution obtained in the simulations seems to have some similarities to the ones obtained in reality, images with better resolution are needed to further assess this aspect.