R Hunt1, R M Roberts1, A M Mortimer2. 1. Department of Neuroradiology, Southmead Hospital, North Bristol NHS Trust, Bristol BS10 5NB, UK. 2. Department of Neuroradiology, Southmead Hospital, North Bristol NHS Trust, Bristol BS10 5NB, UK. Electronic address: alex_mortimer@hotmail.com.
Abstract
AIM: To highlight the magnetic resonance imaging (MRI) signs associated with spinal dural arteriovenous fistula (SDAVF) through categorisation of sources of radiological error and investigation of the delay to radiological diagnosis. MATERIAL AND METHODS: This was a retrospective, observational study of cases referred to a neuroscience centre over 11 years. All patients who underwent spinal digital subtraction angiography (DSA) with a subsequent diagnosis of SDAVF were identified. Prior imaging was reviewed and compared with the formal reports issued. RESULTS: Thirty-seven patients with SDAVF were initially imaged in seven institutions. Radiological abnormalities including intradural vessels (37/37, 100%), cord T2 signal change (34/37, 91.9%), and cord expansion (26/37, 70.2%) were present on prior MRI. These signs were not recognised in 22/37 (59.5%), 7/34 (20.5%), and 15/26 (57.7%) of cases, respectively. Increased T2 signal in the cord was the most commonly identified sign (27/34; 79.4%), but prompted either no diagnosis (7/34; 20.5%) or differential diagnoses including ischaemic, inflammatory, or neoplastic aetiologies or a syrinx in 11/34 (32.4%). An appropriate diagnosis was made on initial MRI in 15 patients (40.5%). The time from initial imaging to diagnosis was significantly delayed for those patients who did not have an arteriovenous vascular aetiology included in the initial differential diagnosis (281 (423.3) days versus 22 (15.7) days, p=0.03). CONCLUSION: SDAVF have imaging features that are frequently missed or misinterpreted. This results in a significant delay to definitive diagnosis and therefore treatment.
AIM: To highlight the magnetic resonance imaging (MRI) signs associated with spinal dural arteriovenous fistula (SDAVF) through categorisation of sources of radiological error and investigation of the delay to radiological diagnosis. MATERIAL AND METHODS: This was a retrospective, observational study of cases referred to a neuroscience centre over 11 years. All patients who underwent spinal digital subtraction angiography (DSA) with a subsequent diagnosis of SDAVF were identified. Prior imaging was reviewed and compared with the formal reports issued. RESULTS: Thirty-seven patients with SDAVF were initially imaged in seven institutions. Radiological abnormalities including intradural vessels (37/37, 100%), cord T2 signal change (34/37, 91.9%), and cord expansion (26/37, 70.2%) were present on prior MRI. These signs were not recognised in 22/37 (59.5%), 7/34 (20.5%), and 15/26 (57.7%) of cases, respectively. Increased T2 signal in the cord was the most commonly identified sign (27/34; 79.4%), but prompted either no diagnosis (7/34; 20.5%) or differential diagnoses including ischaemic, inflammatory, or neoplastic aetiologies or a syrinx in 11/34 (32.4%). An appropriate diagnosis was made on initial MRI in 15 patients (40.5%). The time from initial imaging to diagnosis was significantly delayed for those patients who did not have an arteriovenous vascular aetiology included in the initial differential diagnosis (281 (423.3) days versus 22 (15.7) days, p=0.03). CONCLUSION:SDAVF have imaging features that are frequently missed or misinterpreted. This results in a significant delay to definitive diagnosis and therefore treatment.