Temperature mapping for MR-guided radiofrequency ablation - How accurate can cell necrosis be monitored?

Abstract

Purpose of this study was to quantitatively evaluate the ability of MR temperature measurement performed after thermoablative therapy to predict cell necrosis. Ten patients with primary and secondary liver lesions were treated by MR-guided radiofrequency (RF) ablation. A gradient echo sequence was used to calculate temperature maps (n=22) in the plane of the RF applicators (n=10) and in a plane perpendicular to them (n=12), based on the proton resonance frequency shift method. Slice-identical contrast-enhanced images acquired one month after therapy were registered on the temperature maps and the necrotic, non-enhanced area was segmented and compared to the area with a displayed temperature above 60° C. Sensitivity and Positive Predictive Value of the temperature map was calculated, using the follow-up imaging as gold standard. A subgroup analysis for temperature maps in the plane of the applicators was performed. Temperature mapping reached acceptable image quality in 20/22 cases. Sensitivity, i.e. the rate of correctly detected coagulated tissue was 0.82 ± 0.08. Positive Predictive Value, i.e. the rate of voxels in the temperature map over 60° C that actually developed necrosis was 0.90 ± 0.07. The Positive Predictive Value decreased to 0.87 ± 0.05 for temperature maps in the plane of the applicators, whereas the sensitivity remained unchanged. It can be concluded that MR temperature mapping allows for the prediction of tissue necrosis after radiofrequency therapy with an acceptable accuracy.