MRI has limited agreement with CT in the evaluation of vertebral fractures of the canine trauma patient. Issue 5 (15th July 2019)

• Record Type: Journal Article
• Authors: Gallastegui, Aitor; Davies, Emma; Zwingenberger, Allison L.; Nykamp, Stephanie; Rishniw, Mark; Johnson, Philippa J.
• Is Part Of: Veterinary radiology & ultrasound. Volume 60:Issue 5(2019)
• Page Start: 533
• Page End: 542
• Publication Date: 2019-07-15
• Subjects: accuracy -- computed tomography -- CT -- interobserver agreement -- sensitivity -- specificity -- spine fracture; Veterinary radiology -- Periodicals; Radiologie vétérinaire -- Périodiques; Échographie vétérinaire -- Périodiques; Radiography -- Periodicals; Ultrasonography -- Periodicals; Diagnosis -- Periodicals; Veterinary Medicine -- Periodicals; Societies, Medical -- Periodicals; 636.089607
• DOI: 10.1111/vru.12785
• Languages: English
• ISSNs: 1058-8183

Abstract

Abstract: Complete assessment of vertebral trauma in dogs currently requires CT and MRI for evaluation of the osseous and soft tissue structures that contribute to vertebral stability. Some studies in people have suggested that MRI may be sensitive and specific at detecting vertebral fractures making this potentially a single modality that could be used in spinal trauma evaluation. This study aimed to assess the ability for observers to evaluate vertebral fractures using MRI when compared to CT, which was used as the reference standard. Twenty‐nine dogs with previously diagnosed acute vertebral fractures and four dogs with no vertebral fracture that had undergone sequential CT and MRI were included into the study. One hundred twenty‐eight vertebrae were evaluated for the presence of fractures. Imaging studies were read by two observers blinded to the history. While both observers had similarly high sensitivity and specificity for simple detection of any fractured vertebrae, interobserver agreement was only moderate (κ = 0.584). When evaluations were specifically limited to detection of structurally unstable fractured vertebrae both observers showed improved specificity and interobserver agreement became substantial (κ = 0.650). Complete agreement for exact fracture location between MRI and CT results was only achieved in 14.3‐32.6% of fractured vertebra with up to 79% of fractures being missed in some vertebral structures. This suggests that although MRI may be able to detect the presence of fractured vertebrae, it is not able to replace CT for the complete evaluation of the traumatized spine and documentation of fracture morphology.