Background: The purpose of the present study is to evaluate the accuracy of the thermal dose (TD), calculated from real-time temperature imaging simultaneously to radiofrequency (RF) ablation on pig liver with a clinical device, as an indicator of the induced necrosis.

Methods: The animals (n=4) were maintained under general anaesthesia and positioned inside a 1.5T MR system. An internally cooled MR compatible RF needle was inserted in the liver (14 ablations in total). Electrical interferences of the generator (200W max output power) were suppressed by inserting 2 rejection filters (total attenuation higher than 100dB) tuned to the proton resonant frequency (64 MHz) in the transmission line. A standard RF ablation protocol (12 minutes) was performed simultaneously to continuous MRI acquisition of segmented-EPI gradient echo images. Temperature images were calculated on-line from dynamic phase images using the proton resonance frequency shift method to compute the cumulative thermal dose using Sapareto equation. After completion of the ablations, each animal was sacrificed to take and fix the liver. After two-weeks-fixation, livers were sliced and the largest diameter perpendicular to the electrode axis was measured and correlated to the dimension observed on the corresponding thermal dose maps.

Results: High quality temperature maps were obtained, with a 1.6°C standard deviation of temperature observed in a non heated region, with identical values before and during RF application. No residual artifact due to electromagnetic interferences was observed, except a 10% loss in SNR during RF deposition. The artifact induced by the needle was15±2 mm. An excellent correlation (r=0.98) between TD and histological measurement was found (fig.1). Microscopic analysis of tissue after HES coloration confirmed complete necrosis in the heated region.

Conclusions: This study demonstrates that temperature maps of excellent precision can be obtained on pig liver in vivo on a clinical scanner simultaneously to RF ablation, with the use of appropriate filtering. Apparent lesion dimensions measured on TD maps are in good agreement with the gold standard of histology.