RESUMO
BACKGROUND: One of the current shortcomings of radiofrequency (RF) tumor ablation is its limited performance in regions close to large blood vessels, resulting in high recurrence rates at these locations. Computer models have been used to determine tissue temperatures during tumor ablation procedures. To simulate large vessels, either constant wall temperature or constant convective heat transfer coefficient (h) have been assumed at the vessel surface to simulate convection. However, the actual distribution of the temperature on the vessel wall is non-uniform and time-varying, and this feature makes the convective coefficient variable. METHODS: This paper presents a realistic time-varying model in which h is a function of the temperature distribution at the vessel wall. The finite-element method (FEM) was employed in order to model RF hepatic ablation. Two geometrical configurations were investigated. The RF electrode was placed at distances of 1 and 5 mm from a large vessel (10 mm diameter). RESULTS: When the ablation procedure takes longer than 1-2 min, the attained coagulation zone obtained with both time-varying h and constant h does not differ significantly. However, for short duration ablation (5-10 s) and when the electrode is 1 mm away from the vessel, the use of constant h can lead to errors as high as 20% in the estimation of the coagulation zone. CONCLUSION: For tumor ablation procedures typically lasting at least 5 min, this study shows that modeling the heat sink effect of large vessels by applying constant h as a boundary condition will yield precise results while reducing computational complexity. However, for other thermal therapies with shorter treatment using a time-varying h may be necessary.
