Background: Diagnosis of multiple sclerosis (MS) frequently relies on MRI dissemination in time (DIT) and space (DIS), as codified in 2017 McDonald criteria (McD 2017). The central vein sign (CVS) is a proposed MS diagnostic biomarker, but its optimal incorporation into McD 2017 has not been extensively studied.

Objective: Evaluate the diagnostic performance of several methods incorporating CVS into McD 2017 radiological DIS criteria.

Methods: Data were obtained from the CAVS-MS Pilot, a cross-sectional, international multi-center study conducted by the North American Imaging in MS Cooperative (NAIMS) that recruited adults referred for suspicion/diagnosis of demyelinating disease. Diagnostic performance of methods incorporating CVS into McD 2017 radiological DIS were evaluated by comparing sensitivity, specificity, and accuracy.

Results: 78 participants (37 MS, 41 others) were included. For MS diagnosis, sensitivity, specificity, and accuracy of DIS based on brain imaging (DIS-B) alone was 92 %, 69 %, and 78 %. Requiring at least one lesion with CVS in any brain location in addition to DIS-B increased specificity (sensitivity 92 %, specificity 81 %, accuracy 86 %). Presence of 2 deep white matter lesions with CVS as an additional topography for DIS-B had higher sensitivity (sensitivity 97 %, specificity 59 %, accuracy 77 %).

Conclusions: Incorporation of CVS in McD 2017 DIS criteria can be used to improve diagnostic accuracy. Validation in additional prospective studies is needed.