Integrated ultrasound thermometry and multiphysics modeling for liver RF ablation monitoring: Ex vivo studies

Monitoring of radiofrequency ablation (RFA) is essential to ensure accurate treatment coverage of large liver tumors. Ultrasound is an excellent option for monitoring of ablations. Since ultrasound B-mode imaging lacks the ability to precisely estimate the extent of the lesion, there has been considerable effort and interest in other ultrasound based monitoring techniques. Ultrasound thermometry is one such technique that has the potential to monitor thermal ablations. However, drawback of the approach is its applicability only in sub-ablative temperatures. We attempt to overcome this limitation by integrating a predictive thermal model with ultrasound thermometry and extend its applicability to ablative temperatures. In this paper, we validate this integrated approach through ex vivo studies on bovine liver tissue. Specifically, sub-ablative temperature rise is measured using ultrasound thermometry in a plane distal to the metallic RFA tine. Independently, temperature distribution in the tissue is obtained from a COMSOL™-based multiphysics model (based on combining the Laplace and bio-heat transfer (BHTE) equation) customized for the ablation device geometry and boundary conditions. Subsequently, the experimentally determined ultrasound temperature estimates and the model temperature estimates are compared to estimate the local in situ unknown tissue properties. Finally, the adapted model is utilized during the ablation to estimate the ablation size. The thermal and electrical conductivity estimated using our approach for N=5 samples are 0.59±0.07 W/m/°C and 0.20±0.02 S/m respectively. Further, the lateral and axial lesion widths are 14.6 ±2.3 mm and 30.6±1.1 mm respectively and compare well with gross pathology estimates of 12.6±1.8 mm and 25.6±1.3 mm. In ex-vivo conditions devoid of blood perfusion effects, we have demonstrated the successful integration of the multiphysics model with the experimentally measured ultrasound backscatter data and the feasibility of the approach to monitor ablations.