The future of preoperative planning for liver cancer : from CT-scan in the 20th century to computational model in the 21st century

Since the development of CT-imaging in the 20th century, medical imaging has transformed the field of preoperative surgical planning. Today, 3D reconstructions of patient-specific anatomies are increasingly being used as input for computational models of various treatments. These treatment models can assist the clinician in comparing different treatment scenarios preoperatively and choosing the most optimal one for each patient-specific situation, thereby increasing health outcomes. Specifically, for liver cancer, we have developed a computational fluid dynamics (CFD) model in Fluent (Ansys, USA) of blood ($\rho$: 1060 kg/m³) and drug particle ($\rho$: 1060 kg/m³, diameter: 40 µm) flow in the hepatic arteries perfusing the tumor, based on the conebeam CT images of a patient with primary liver cancer treated at UZ Leuven. This CFD model allows us to predict the impact of various interventional parameters on the tumor dose distribution. For example, a microcatheter (length: 80 mm, diameter: 0.7 mm) positioned near the wall led to a much more uniform dose distribution than a microcatheter positioned at the center of the vessel. However, the total dose delivered to the tumor was only 43.2\% for the off-center catheter, and 49.3\% for the central position. These results show that different microcatheter positions can have a significant impact on both tumor dose and tumor dose distribution. Success stories of preoperative planning software companies like HeartFlow and FEops have shown the incredible potential of computational models in medicine. However, many of these computational models, including our own, need to be validated for a wide range of patients first before being integrated into clinical practice. Additionally, high computational complexity is another factor that might hamper clinical integration. In the near future, as medical imaging has done in the previous century, computational modelling might be the next big step in transforming the field of preoperative surgical planning