Computed tomographic cervical myelography in horses: Technique and findings in 51 clinical cases

Abstract Background Three‐dimensional computed tomographic (CT) evaluation of the cervical vertebral column enables more accurate identification of osseous and soft tissue lesions than traditional latero‐lateral radiography. However, examination of the complete cervical vertebral column has been limited by horse size, preventing evaluation of the caudal cervical vertebrae. Objectives To describe a technique to enable CT myelography of the complete cervical spine and describe the findings in 51 horses. Animals Records of 51 horses presented for evaluation of cervical vertebral lesions. Methods A retrospective review of clinical records from all horses presented for CT myelography to further investigate possible cervical vertebral lesions was performed. A description of a novel approach to CT myelography in horses and retrospective review of the findings in clinical cases has been included. Results Degenerative joint disease was identified at 1 or more dorsal articular process joint in 50/51 horses, of which 44/51 had a site of grade 2 or greater. Spinal cord compression was observed on CT myelography in 31/51 horses, whereas attenuation of the dorsal contrast column was identified radiographically in 11/50 horses. Thirty‐three horses showed narrowing or obliteration of the intervertebral foramina at 1 or more site and osteochondral fragments were seen in 11/51 horses. Conclusions and Clinical Importance Computed tomography myelography is relatively safe and an easily performed technique with the correct equipment, enabling evaluation of the cervical vertebral structures of horses in all planes and volumetrically. It is possible that lesion extent might be underestimated with this diagnostic modality, hence interpretation should be complimented with flexed and extended views radiographically.


| INTRODUCTION
Historically, evaluation of the cervical vertebral column for evidence of vertebral canal stenosis, spinal cord compression (SCC), and degenerative joint disease (DJD) of the articular process joints (APJs) in equids has relied on radiographs and radiographic myelography. [1][2][3][4] Survey radiography has a sensitivity of 42 and 63% for vertebral canal stenosis identified with intravertebral ratio (IVR) and DJD, respectively, when compared with necropsy, and radiographic myelography (judged by 50% reduction in the dorsal contrast column [DCC]) has a sensitivity of 71 and 43% when compared with necropsy. 5 Although radiographic myelography enables assessment of SCC by measurement of contrast attenuation in flexed, neutral and extended positions, and measurement of both intervertebral and intravertebral ratios, threedimensional examination is not possible. 1,6,7 Misdiagnosis of compressive lesions with survey radiography alone occurs in 30%-60% of cases compared with radiographic myelography, 8 and 61% of cases compared with necropsy, 5 highlighting the importance of myelography for evaluation of compressive lesions. Similarly, accurate evaluation of the impact of DJD of the APJ on SCC is limited. 5,6,8 However, in a multicenter study, radiographic myelography had a 68% correlation with necropsy findings, 5 indicating that radiographic myelography frequently underestimates presence of cervical SCC. Computed tomography (CT) and CT myelography are advocated for tomographic evaluation of the cervical vertebral column and identification of SCC. [9][10][11] CT and CT myelography confers multiple advantages including avoidance of superimposition of adjacent structures enabling accurate identification of lesions circumferentially, multiplanar image reconstruction and examination for SCC in any image plane. This might facilitate diagnosis of articular process impingement on spinal nerves as well as SCC.
The use of a specifically designed and adapted CT gantry to enable CT myelography under general anesthesia and the findings of CT myelography in both cadavers and live horses has been reported. 11,12 Compressive spinal cord lesions confirmed histologically were more likely to be accurately detected using CT over radiographic techniques, with fewer false-positive diagnoses. 12 This study describes in detail a technique for CT and CT myelography of the cervical and cranial thoracic vertebral canal in clinical cases, and the CT and CT myelographic findings in 51 horses in which cervical SCC was considered a differential diagnosis.

| MATERIALS AND METHODS OF THE NOVEL CT MYELOGRAPHY TECHNIQUE
A thorough neurological examination was performed in all cases, and was considered indicated for neurolocalization and identification of abnormalities that might influence the safety of general anesthesia.
Standing survey radiographs were performed if a fracture or subluxation of the cervical vertebrae was suspected. Before induction of general anesthesia, an area measuring 20 cm × 20 cm caudally to the occipital protuberance and equilaterally from the midline was clipped.
A catheter was aseptically placed in the right jugular vein and sutured in position. Horses received 0.02 mg/kg acepromazine (Acesedate,  Figure 1A,B). A cushion was positioned under the lower left forelimb to prevent radial nerve paralysis. The forelimbs were extended caudally and fixed in this position with ropes that passed through the hindlimbs and were secured to metal hooks in the table. This positioning enabled image acquisition from the brain to the 7th cervical vertebrae in most horses in 1 continuous study, and the third thoracic vertebrae in some horses. Caudal extent of scanning was most dependent on optimal horse positioning and the thickness of the horse from sternum to withers obstructing the horse fitting within the CT gantry.
During acquisition of the plain CT, the head and neck were positioned flat and slightly extended on a soft pad. For cerebrospinal fluid (CSF) removal and subsequent contrast administration, the horses' head and neck were elevated to a 30 to 45 angle using a custom made wooden triangular block. The previously clipped area was aseptically prepared and a sterile iodinated self-adhesive drape applied to the area.
Puncture of the subarachnoid space at the atlanto-occipital joint was done in accordance with reported techniques. 13  Iohexol 300 mg I/mL with 30 mL Hartmann's solution) over a total timed 180 second period; a total of 80 mL of injectate into the subarachnoid space. Using this novel approach, the initial non-diluted contrast media was "chased" by subsequent diluted contrast media subjectively reducing the ventral pooling effect that occurred with standard protocols. A bolus of CSF was then used solely to flush the remaining contrast in the extension set into to the subarachnoid space and the needle was withdrawn.
The head remained elevated after injection for 5-minutes before being returned to a horizontal position, after which CT image acquisition was performed from the caudal cervical region to the brain.
Subsequently horses were moved into the anesthesia recovery box and a series of laterolateral radiographic projections were obtained in extended and maximally flexed positions of the cranial, mid and caudal cervical regions to complete the dynamic myelographic study.
After complete image acquisition the head and neck were maintained in an elevated position for a further 10 minutes to encourage caudal flow of iohexol (Omnipaque, GE Healthcare AS), and the horses were recovered from anesthesia using a rope recovery system in accordance with the hospital anesthetic protocol.  (Table 2). Radiographic myelogram findings were considered positive for SCC if there was 50% narrowing of the dorsal and ventral contrast column or 20% narrowing of the dura at any site with the exception of C7-T1 for which narrowing of 60% of the contrast column or 30% narrowing of the dura or more was used, based on the findings of a recent study by Estell et al. 14 Anesthetic recovery was graded from 1 to 5 based on previously published scoring systems. 15 The presenting clinical signs, findings of neurological examination, documented adverse events and short-term outcome were reviewed retrospectively and have been summarized in Table 1 with respect to the associated CT and myelographic findings. Short-term outcome was defined as discharge from hospital.

| RESULTS
Fifty-one cases were reviewed. Of these, 17 were mares and 34 were geldings. The median age of horses was 7 years with a range of 1 to 26 years, although the age was unknown in 9 horses. Breeds included  Table 1.     is typically limited to a sagittal plane. 1,6,7,11 In a study of 306 ataxic horses, cervical radiography was unreliable, with an accuracy of 40% for diagnosis of SCC when compared with radiographic myelography. 8 However, even radiographic myelography is limited to evaluation of dorsoventral compression, with lateralized compression requiring dorsoventral radiographic projections which provide inadequate tissue contrast and detail. 8 In that study, SCC was diagnosed when there was more than 50% reduction in both the dorsal and ventral contrast column, which supports other reports that a 50% reduction in contrast column dorsally or ventrally is not a reliable indicator for compressive lesions and should be interpreted in line with both intervertebral and intravertebral sagittal ratios to improve the accuracy. 3,6 In addition, site and position (neutral compared with flexed or extended) specific cutoff values have been suggested, 16 and are appropriate for the cervico-thoracic articulation. 14 The finding of more than 50% reduction in DCC at C7-T1 in normal horses has prompted the suggestion that the cutoff at this site should be increased to more than 60% reduction in the DCC, or more than 20% narrowing of the dural diameter. 14  The majority of grades 2 and 3 lesions were localized to C4-C5, C5-C6, and C6-C7, with grade 3 or severe lesions primarily localized to C5-C6 and C6-C7, consistent with previously reported findings. 7,8,22 Although CVSM is frequency associated with lesions at C3-C4, the primary cause of SCC in the current study was DJD of the APJs, which is similar to other reports 11,22  The risk of general anesthesia in horses with neurological dysfunction or a history of trauma should be considered. General anesthesia in horses with severe neurological dysfunction that have not been observed to lay down/stand up should be approached with caution, and survey radiography is indicated to identify the presence of fractures or subluxations that might further complicate recovery from anesthesia or require surgical interventions. Adverse effects of intrathecal contrast media administration are rarely reported in horses. 23,24 The authors have found this procedure to be relatively safe with an incidence of anesthetic recovery 25  reported. 24 Neurological adverse reactions including seizures, worsening of neurologic grade, somnolence, head shaking, and hyperesthesia were noted most frequently in that study, which is consistent with the literature. 23,24 The incidence of complications was similar in this study, with 5% of cases having a grade 1 or 2 recovery and total morbidity after the procedure of 36% with none showing clinical signs necessitating euthanasia and no horses displaying seizure like activity, hyperesthesia or head shaking. One horse sustained a catastrophic fracture during recovery from general anesthesia making an incidence of 2% in a relatively small sample size, which is similar to the incidence of this occurrence for horses recovering from general anesthesia for any purpose. 25 Intrathecal administration of iohexol is associated with mild irritation and extradural oedema, often resulting in CSF leucocytosis, although it has been reported that the use of a 300 mg I/mL as used in this study was associated with less irritation than a 350 mg I/mL solution. 23,26 Increased speed of contrast delivery, increased volume of contrast media delivered, and increased total anesthesia time have been associated with increased risk of adverse neurological events. 24 However, given the often compromised nature of these horses' proprioception, the incidence of complications being similar to the literature for normal horses supports the relative safety of myelography. This might be in part because of the relatively short anesthetic duration, use of a rope recovery system in horses deemed to be high risk, use of total IV anesthesia and the administration of anti-inflammatories preoperatively. In the authors' opinion, the risks of anesthetic recovery are outweighed by the superior diagnostic and prognostic information provided by CT and CT myelography.