Early detection with MRI of incomplete treatment of spine metastases after percutaneous cryoablation.

To evaluate post-ablation MRI for the detection of incompletely treated spinal osseous metastases (SOM) after cryoablation and to propose a post-ablation imaging classification. After IRB consent, all patients treated with cryoablation of SOM between 2011 and 2017 having at least 1-year minimum follow-up and a spine MRI within 4 months after cryoablation were retrospectively included. A classification of MRI images into four types was set up. The primary endpoint of our study was to assess the diagnostic performance of the post-ablation MRI. The secondary endpoints were the 1-year complete treatment rate (CTR) and complications. Fifty-four SOMs in 39 patients were evaluated. Post-ablation MRI was performed with a median delay of 25 days after cryoablation. Images were evaluated by two independent readers according to the pre-established image classification. Sensitivity and specificity for the detection of residual tumor were 77.3% (95%CI = 62.2-88.5) and 85.9% (95%CI = 75.0-93.4), respectively. Types I, II, III, and IV of the classification were associated with a 1-year complete treatment in 100%, 83.3%, 35.7%, and 10% of cases, respectively. The 1-year CTR was 59.3% for all 54 metastases, and 95.8% for metastases measuring less than 25 mm and at least 2 mm or more away from the spinal canal. Two grade 3 and two grade 2 adverse events according to the CTCAE were reported. MRI after cryoablation is useful for the evaluation of the ablation efficacy. The classification of post-cryoablation MRI provides reliable clues for the prediction of complete treatment at 1 year. • MRI performed 25 days after cryoablation is useful to evaluate the efficacy. • The proposed classification provides a reliable clue for complete cryoablation. • Percutaneous cryoablation of spinal metastases is highly effective for lesions less than 25 mm in diameter and of at least 2 mm away from the spinal canal.