Electro-physiological parameters of hepatic radiofrequency ablation—a comparison of an in vitro versus an in vivo porcine liver model

Abstract

Background

Radiofrequency ablation (RFA) is an inherent part of curative treatment within a multimodal therapy concept of malignant liver tumors. The biggest problem is the high rate of local recurrences in tumors with a diameter of more than 3 cm because of the high variability and poor reproducibility of the zone of ablation. No imaging modality facilitates monitoring during neither intraoperativ nor percutaneous RFA. This experimental study describes and compares an in vitro and in vivo porcine model by its electro-physiological parameters with the aim of monitoring RFA procedures.

Materials and methods

RFA was performed in a perfused in vitro porcine (one RFA per liver) and in vivo porcine model (24 animals) with three different RFA systems (Rita XL 5 cm, Rita XLi 7 cm, LeVeen 5 cm). In the in vivo model, percutaneous placement of the RFA device was guided by native and contrast-enhanced CT scan. The electro-physical parameters during RFA were online (in real time) recorded by a dedicated software. After the RFA, the livers were explanted, sliced, and measured according to the consensus technique.

Results

The delivered energy was in vivo versus in vitro: Rita XL 238 ± 135 kJ versus 135 ± 53 kJ (p = 0.247); Rita XLi 711 ± 180 kJ versus 159 ± 54 (p = 0.016) and with LeVeen 212 ± 71 kJ (in vivo). The LeVeen system was inconsistent in the in vitro model. This correlates to an energy consumption per ml of necrosis in vivo versus in vitro Rita XL of 8 ± 3 kJ/ml versus 6.4 ± 3.9 kJ/ml (p = 0.537), Rita XLi of 10 ± 6 kJ/ml versus 1.8 ± 0.2 kJ/ml (p = 0.016), and LeVeen of 14.0 ± 12 kJ/ml (in vivo). The volume of ablation was in vivo versus in vitro Rita XL 30 ± 10 ml versus 26 ± 17 ml (p = 0.329), Rita XLi 90 ± 58 ml versus 88 ± 21 ml (p = 0.905), and LeVeen 22 ± 11 ml versus 50 ± 12 ml (p = 0.04). The impedance during RFA were in vivo versus in vitro Rita XL 39 ± 4 Ω versus 50 ± 14 Ω (p < 0.247), Rita XLi 33 ± 5 Ω versus 61 ± 16 Ω (p = 0.016) and LeVeen 31 ± 2 Ω (in vivo).

Conclusion

The volume of ablation showed analogue data in vivo and in vitro. The delivered energy and energy consumption was in vivo up to five times (Rita XLi) higher than in vitro and the impedance in vivo was always lower than in vitro. These differences observed between in vivo and in vitro were more pronounced than previously described. Thus the use of an in vitro model for research of the RFA technique must be challenged. The large deployment of the Rita XLi was a problem for percutaneous positioning of the device without direct contact to liver surface or major vessels in 80-kg pigs and to a lesser extent in in vitro liver originating from 130- to 140-kg pigs. Modern RFA systems which generate large volume of tissue necrosis can therefore only be adequately tested in a porcine model with a liver weight of at least 1.5–2 kg. Alternatively, a bovine liver model (with a liver weight up to 10 kg) should be developed in the future.