Table 6.

Factors that influence the ability of make an accurate assessment of CM and SM on MRI.

Type	Example	Notes
Protocol	Slice thickness	In the sagittal plane thinner slices (3 mm or less) are preferred to achieve 2–3 sections through the spinal canal and more chance of lesion detection. Thicker slice thickness may miss a small intramedullary lesion in an dog with a spinal cord diameter ranging between 4.1 and 10.3 mm (depending on site imaged and size of animal) (107, 108). However, MRI machines of 1 Tesla or less cannot provide thin slices with sufficient signal-to-noise ratio. Using a balanced steady state feed precession sequence (bSSFP) in addition to a conventional gradient echo and spin echo sequences may help overcome this challenge (109).
	Sequence	Protocols that achieve both T1W and T2W weighting are required to be confident in detecting a fluid filled cavity within the spinal cord and conformational changes with CM.Anatomical imaging of the entire brain is recommended and full extent of the syrinx should be determinedTransverse images perpendicular to the spinal cord though the syrinx are required to assess the transverse width and extent of spinal cord involvement.
Magnetic field strength	Low field vs. high field	Signal-to-noise ratio and spatial resolution is improved when imaging with higher magnetic field-strength which allows shorter imaging times for a given resolution and/or higher resolution for a given imaging time. In addition, higher signal-to-noise ratio allows better resolution with smaller voxel size and thinner slice thickness (110).
Operator factors	Inexperience/lack of training	In veterinary medicine it is possible to operate a MRI service without any Specialist qualification. By contrast an experience MRI technician has undertaken a 3–4 year radiography degree plus additional post-graduate MRI training.
	Diligence	Out with other reasons for decreasing imaging time (economic/duration of anesthesia), operator inclination is a factor for example image quality can be improved by increasing the number of averages (NEX/NSA) which will subsequently increase the acquisition time.
Interpreter factors	Inexperience/lack of training	Failure to recognize significant lesions or over-interpretation of other features for example attributing SM to epilepsy, facial nerve paralysis, fly catching and other brain disorders or interpreting a generalized pruritus as due to SM.In humans it is reported more likely that a cervical syrinx is missed with techniques for whole spine sagittal scanning with focused lumbar spinal MRI where the physician is biased from the history for a lumbar lesion (111). Conversely, asymptomatic localized widening of the central canal may be observed in both humans and dogs (34, 112).
Patient factors	Skull and air interface	May cause susceptibility artifacts, especially on gradient echo sequences.
	Small brain and narrow spinal cord	Slice thickness should be proportional to the brain volume to achieve images with diagnostic quality i.e., animals with smaller brain volume require thinner slices. In a low field MRI this may be challenging and a bSSFP sequence is recommended.
	Positioning	Assessment for CM is normally obtained with the head in extension as reproducibility is easier and anesthesia is safer as the airway may be compromised in the flexed position (33). However, cerebellar herniation and CSF space between the cerebellum and brainstem are significantly increased in the flexed position (113).
	Microchips, orthopedic implants and shrapnel	Ferromagnetic materials cause susceptibility artifacts which may compromise interpretation especially identity microchips for studies of the cervical spine. In low field MRI a T1W turbospin echo sequence is recommended (114) and for high field MRI, spin echo sequences have smaller artifacts than gradient echo sequences (115). Titanium or oxidized zirconium implants have less susceptibly artifact than cobalt-chromium alloy implants (116)
	General anesthetic	Increased time under general anesthesia may increase risk to patient and cost thus limiting length of any MRI protocol.
Motion related artifacts	Neural tissue	Standard MRI sequences are optimized for good spatial and contrast resolution, however this results in blurring of moving structures which compromises the ability to detect fine structures such as arachnoid webs and other adhesions, septations in the syrinx or appreciate dynamic compression (104). It may also blur the edges of a syrinx cavity.
	Intrasyringeal fluid flow void	Pulsatile or turbulent motion of fluid within the syrinx produces low signal on T2W images because of an absence of activated protons in that region (117)
	Intraventricular CSF pulsation artifact	Intraventricular hyperintensity on FLAIR imaging which result in false—negative/positive interpretations of ventricular pathology and is a particular problem for FLAIR performed on low field MRI (118). The most common cause is pulsatile movement and un-inverted CSF flowing into the slice between the pulses (119). It can also occur because of inadequate inversion of CSF magnetization at the edge of the transmitted coil or because of increased CSF protein or oxygen (breathing 100% oxygen) which shortens T1 (119).