Unenhanced magnetic resonance imaging immediately after radiofrequency ablation of liver malignancy: preliminary results

Purpose: To assess the accuracy of unenhanced magnetic resonance imaging (MRI) immediately after the percutaneous ultrasound-guided radiofrequency ablation (RFA) of liver malignancy in predicting treatment efficacy at CT follow-up. Materials and methods: Percutaneous ablation was prospectively performed in 23 liver malignancies (20 hepatocarcinomas and 3 metastases). After the procedure in the same day all patients were studied with unenhanced MRI. The best sequence to detect the coagulative necrosis was visually established. Pre-RFA CT and post-RFA MRI were registered with non-rigid transformation algorithm. Manual segmentation of lesions and ablated areas in pre-RFA CT, post-RFA MRI, and follow-up CT were obtained. Sensitivity, specificity, positive predictive value (PPV), negative predicitve value (NPV), and accuracy of MRI in predicting the correct centering and the complete treatment of the lesion were calculated in respect to the 1-month follow-up CT. Results: Fat-saturated T1-weighted (fs T1-w) was the sequence in which the best conspicuity of the ablated area was depicted. Coagulative necrosis was hyperintense in fs T1-w sequence in 17/23 (74%). In respect to follow-up CT, MRI predicted the correct centering of the lesions in 19/20 lesions with sensitivity, specificity, PPV, NPV, and accuracy of 100%, 75%, 95%, 100%, and 100%, respectively. MRI predicted the complete treatment of the lesions in 17/17 lesions with sensitivity, specificity, PPV, NPV, and accuracy of 100%. Conclusion: MRI with the single fs T1-w sequence was highly accurate in predicting the treatment efficacy of percutaneous ablation of liver malignancies in comparison to follow-up CT control. Unnecessary CT in case of incomplete treatment can be therefore easily avoided.