Modeling and simulation of thermal ablation in vascularized tissues
MetadataShow full item record
Thermal ablation, which uses localized heating by external energy sour-ces to destroy abnormal tissues, has been increasingly used as a treatment modality for cancers. Its efficacy hinges on the ability to control temperature and induce sufficient thermal damage in the targeted tissue regions. However, it is difficult for clinicians to decide what treatment settings should be applied to obtain the best treatment outcome while minimizing adverse effects. Help from numerical simulations in the treatment planning process would therefore be highly valuable, if the predictions are realistic, reliable and accurate. The complexity of tissue structures and the coupling of heat transfer with other biophysical phenomena such as blood perfusion, tissue damage and thermoregulation make it a challenge to predict temperature fields and thermal damage in living tissues. Approximate models with uncertain parameters are then used in the simulations. The important question is, which features are necessary to include in a model for plausible predictions to guide clinical treatment? Furthermore, what can be done to improve prediction accuracy? To address these issues a series of numerical simulations of radiofrequency (RF) ablation were performed that include the following effects: 1) the thermal effects of micro vascular perfusion, 2) the impact of tissue damage on that perfusion, and 3) the thermal effects of discrete blood vessels. In addition two distinct models of the thermal effects of perfusion, with very different underlying assumptions were used: Penne's bioheat transfer model, and a porous media model. The former is expected to be valid at large scales (e.g.,~the scale of a whole organ), while the latter is expected to be valid at small scales (e.g.,~the scale of a capillary bed). It is not clear what perfusion model is most appropriate at the intermediate scale of thermal ablation. The results of these simulations were analyzed to identify the most important factors to consider in treatment planning for RF ablation. By far the most important uncertainty in predicting the tissue damaged by RF ablation therapy was due to the difference between the two ablation models studied, resulting in a factor of two difference in the predicted volume of damaged tissue. Also important were the effective porosity of the tissue and the perfusion rate, especially in the case of the Pennes' perfusion model, and the presence of blood vessels larger than 0.5 mm in diameter and closer than 20 mm from the applicator. Of particular interest is the finding that it is very difficult to damage all tissue very close to a blood vessel larger than 3 mm in diameter, because of the cooling it provides. These observations suggest improved perfusion modeling, reliable determination of effective porosity and perfusion rate, and the mapping of large vessels in the treatment area would all improve the ability to predict the response to RF thermal ablation for treatment planning.