Radio-frequency-induced thermoablation: monitoring with T1-weighted and proton-frequency-shift MR imaging in an interventional 0.5-T environment.

Abstract

To evaluate the feasibility and accuracy of monitoring radio-frequency (RF) ablation with an open-configuration, 0.5-T magnetic resonance (MR) imager. Thirty-six in vivo RF ablation experiments were performed in porcine paravertebral muscle (n = 24) and liver (n = 12). A 90 degrees C tip temperature was applied for 3-9 minutes. MR images were acquired after continuous or during intermittent RF application. Temperature changes were monitored as signal intensity and proton-frequency-shift (PFS) alterations in two T1-weighted gradient-echo sequences. An update image was obtained every 2.5 seconds (20/10 [repetition time msec/echo time [TE] msec]) or every 5.0 seconds (40/20). A color-coded subtraction technique enhanced the signal intensity and PFS changes. Macroscopic coagulation size was compared with MR image lesion size. The RF application mode had no significant effect on coagulation size in muscle or liver (P > .05). Twenty-two of 24 coagulative lesions in muscle and nine of 12 in liver were demonstrated with the PFS technique. Accuracy of lesion size determination depended on TE (TE = 20 msec, r = .95; TE = 10 msec, r = .78 [P < .01]). The T1-weighted technique depicted only six of 24 muscle and three of 12 liver lesion ablations. In the lesions depicted, macroscopic size was significantly underestimated (P < .001). PFS MR monitoring of RF effects in liver and muscle is feasible and accurate. The PFS technique outperformed the T1-weighted technique.