Purpose To evaluate the accuracy of MRI for the diagnosis of spinal vascular malformations (SVM).
Materials and methods We reviewed 133 consecutive patients with 142 high-flow SVMs confirmed by spinal angiography between 1999 and 2104. The initial MRI diagnosis was available in 132 instances. MRI was considered positive when a SVM was listed in the differential diagnosis or when spinal angiography was recommended. It was considered falsely negative otherwise (i.e., no or wrong diagnoses).
The lesions included 89 spinal dural or epidural arteriovenous fistulas (SDAVF/SEAVF) (67.4%), 26 perimedullary arteriovenous fistulas (PmAVF) (19.7%), 8 spinal arteriovenous malformations (SAVM) (6.1%), 3 SAVM/PmAVFs (2.3%), and 6 paravertebral arteriovenous fistulas (ParAVF) (4.5%). The PmAVFs were subdivided into 8 Type 1 (6.1%), 15 Type 2 (11.4%), and 3 Type 3 (2.3%).
Results The initial MRI diagnosis was falsely negative in 50% of cases (66 SVMs in 63 patients). Missed lesions included 54 SDAVF/SEAVFs (81.8%) and 12 PmAVFs (18.2%). No case of SAVM, SAVM/PmAVF, or ParAVF was missed. When grouped, slow-flow lesions (54 SDAVF/SEAVFs and 8 PmAVF Type 1) accounted for 93.9% of the false negative MRIs, with a 63.9% false negative ratio (62 missed lesions out of 97).
The most common MRI misdiagnoses included inflammatory conditions (transverse myelitis) (46.0%), spine degenerative disease (11.1%), tumors (7.9%), and stroke (3.2%). No specific diagnosis was offered in 31.7% of cases.
Conclusion The initial MRI in patients with SVMs was falsely negative in half of the reviewed cases. Slow-flow spinal arteriovenous fistulas (SDAVF, DEAVF, and PmAVF Type 1) were particularly insidious lesions, missed in about two-thirds of instances. About half the patients received an incorrect diagnosis of TM and were threated accordingly. Most patients with no specific MRI diagnosis were treated empirically for TM as well.
Missed initial diagnoses not only delay adequate treatment for conditions with narrow therapeutic windows, but also result in the introduction of inappropriate, often deleterious management (e.g., high-dose steroids or plasma exchange). Our findings, based on patients with angiographically confirmed lesions, likely underestimate the true rate of false-negative MRI diagnoses in patients with SVMs.
Disclosures D. Sorte: None. E. Wyse: None. E. Orru': None. P. Gailloud: None.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.