By Dr Poneh Adib-Samii, Consultant Neurologist, Croydon University Hospital

This project was awarded runner up in the MS Academy MasterClass 5 Project Award 2019



Magnetic resonance imaging (MRI) is a cornerstone in the diagnosis and treatment monitoring of MS. Imaging guidelines in MS have been proposed by MRI in MS (MAGNIMS, 2015) & Consortium of MS Centers (CMSC). CMSC (2016) proposed four standard protocols (brain, orbits, spine and PML), updated in 2018 to include the judicious use of gadolinium. The Barnet MS service has over 600 patients mostly on disease modifying therapies (DMT). The MRI departments are extremely busy serving a large community across multiple sites. There are no dedicated neuro-specialised scanners, radiologists or radiographers. The objective of this audit was to compare
acquired MRI in suspected MS and CIS with CMSC guidelines.



Retrospective audit of new referrals to MS Clinic during a one-year period between June 2017 and 2018. Cases of confirmed MS or those on DMT were excluded. This was a cross-sectional study collecting clinical information using CERNER and extracting DICOM information using Carestream.

There were 22 new patients (5 males: 17 females) with a mean age of 33.5 (range 19 to 51). They were scanned on 5 different scanners (all 1.5T) across 4 sites. The following imaging was requested: 3 brain, 12 brain with cervical spine, and 7 brain with whole spine. Post contrast imaging was acquired in 11(50%) brain and 5 (25%) spinal scans.



All brain imaging was of the required pixel resolution of ≤ 1 x 1mm. However, slice thicknesses were typically 5 or 6 mm with a gap (typically 1mm), compared to the recommended ≤ 3mm without a gap. There were at least six different acquisition protocols. Only 3 (14%) had the required sagittal and axial FLAIR sequences. However, 14 (64%) did have axial and sagittal T2. There were no 3D FLAIR or T1 (IR-prep GE T1, PSIR or MPRAGE) images. All cases had 2D T1 with only 15 (68%) axial T1.

All spine imaging was of the required pixel resolution of ≤ 1 x 1mm. 95% had slice thickness of 3mm (recommended ≤ 3mm), however, typically with a gap of 0.3 mm. CMSC recommends two sagittal sequences: T2, PD, STIR or T1-PSIR. 15 (79%) had T2 and STIR. There were at least seven different acquisition protocols. All had T2 axials (usually lumbosacral) but only 60% (6/10) had adequate lesion coverage.



All imaging fells short of recommended guidelines with large slice thicknesses and/or gaps resulting in partial volume effects. Brain sagittal and axial FLAIR (only 14%) is recommended for better detection of juxtacortical & peri-ventricular lesions compared with T2. Most FLAIR was coronal (86%) making axial T2 lesion corroboration difficult. Adequate axial T2 or PD is recommended for better detection of posterior fossa lesions compared to FLAIR. The most common T2 protocol was BLADE, optimized for motion artifact but with poor resolution.

There were no 3D volumetric T1, however, time & suboptimal images at 1.5T may preclude its clinical use. Radiographers fed back that the minimum delay of 5 minutes post contrast was not adhered to, potentially reducing enhancement detection rate.
For spine imaging 80% had T2 and STIR but no PD. PD may be better than STIR for cord signal and can be acquired with T2 using a long repetition time (dual echo).

In summary, MRI at Barnet is non-standardised with variation in hardware, software and protocols. Imaging is acquired across multiple sites within set time slots. Any future protocol must consider these factors, meet all clinical needs, and be agreed with third party private providers.