Diffusion tensor imaging along the perivascular space and disease progression in people with multiple sclerosis: a 5-year longitudinal MRI study

Diffusion Tensor Image Analysis ALong the Perivascular Space (DTI-ALPS) was originally proposed to quantify glymphatic functioning. Although a direct interpretation is now questioned, cross-sectional studies show associations with disability in people with multiple sclerosis (pwMS). Regardless, serial DTI-ALPS studies are largely lacking in MS. In a longitudinal study, we investigated DTI-ALPS with respect to confirmed disability progression (CDP) and progression independent of relapse activity (PIRA) in people with relapsing-remitting MS (pwRRMS) and progressive MS (pwPMS). This study included 72 pwRRMS, 27 pwPMS, and 23 healthy controls (HC) imaged with 3T MRI and again after 5 years. The DTI-ALPS index was calculated using an automated pipeline using template-defined regions of interest (ROIs) in the superior longitudinal fasciculus and superior corona radiata. Areas corresponding to T2 hyperintensities were removed to avoid the influence of overt pathology. CDP and PIRA were assessed after 5 years and in 64 pwRRMS/17 pwPMS after 10 years. Comparisons between those with and without follow-up CDP or PIRA were assessed using analysis of covariance and repeated including normal appearing white matter (NAWM) mean diffusivity (MD) as an additional covariate. Multivariable binary logistic regression was used to explore whether DTI-ALPS offers independent value beyond general disease burden. Although significantly lower in pwMS compared with HCs (1.347 ± 0.178 versus 1.437 ± 0.132, P = 0.034, partial η² = 0.021), the difference was no longer so after controlling for NAWM MD (P = 0.094, partial η² = 0.024). DTI-ALPS decreases over 5 years were similar between HC and pwMS (P = 0.188, partial η² = 0.021). In pwRRMS, baseline DTI-ALPS was lower in those who developed CDP or PIRA at 5- and 10 years of follow-up (all P ≤ 0.019, partial η ² > 0.080, except for PIRA at 5 years, P = 0.051, partial η² = 0.055). When controlling for NAWM MD, results were in line with original findings. Baseline T2-LV was the only retained imaging predictor of CDP and PIRA over 5 years while only baseline DTI-ALPS was selected for in 10 year models. No associations were found in the pwPMS group. Changes in DTI-ALPS over 5 years did not relate to CDP nor PIRA in neither group. In conclusion, although DTI-ALPS values were not significantly different compared with HCs after considering NAWM MD, decreased baseline DTI-ALPS is associated with disability progression in pwRRMS. The lack of associations in pwPMS suggests that DTI-ALPS may be less informative with more advanced disease.