Background

Early detection of sub-optimal treatment response in relapsing-remitting MS is important for effective treatment escalation, but currently no method exists for reliable detection of early poor treatment responders. Previously we demonstrated an association between blood-brain barrier (BBB) permeability, as measured by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), and cerebrospinal fluid cellularity.^1,2 We hypothesized that DCE-MRI predicts sub-optimal treatment response after two years of treatment with fingolimod or natalizumab, drugs with a common effect of decreasing lymphocyte influx into the CNS.

Methods

35 relapsing-remitting multiple sclerosis patients starting on fingolimod or natalizumab were scanned with DCE-MRI at 3T, prior to treatment initiation (baseline) and again three and six-months post treatment. We calculated the influx constant K_i, a measure of BBB permeability, using the Patlak model as previously described^3,4. Sub-optimal treatment response was defined as loss of no evidence of disease activity (NEDA-3) status after two years of treatment. NEDA-3 is defined as absence of disease activity in three domains: 1) no new or enlarging T2 lesions on MRI, 2) no clinical disease activity in the form of relapses, and 3) no increase in disability as measured by the expanded disability status scale (EDSS). Two subjects who did not complete one year of treatment were excluded.

Results

Baseline K_i in normal appearing white matter (NAWM) was significantly predicted by methylprednisolone treatment 2 months prior (β=-0.71, p=0.00008) and days since last relapse (β=-0.48, p=0.005), but not by first-line treatment (yes/no; p=0.55) or visibly contrast enhancing lesions (p=0.76) in a linear regression analysis (model R²=0.40, p=0.0002), see Table 1 and Figure 3. 12 out of 33 (36%) subjects lost NEDA status after one year of treatment and 15 out of 33 (46%) lost NEDA status after two years. Subjects who lost NEDA status at two years had a 51% higher mean K_i in normal-appearing white matter (NAWM) measured after six months of treatment, when compared to the group who maintained NEDA status (p=0.0003), while there was no difference in K_i before treatment initiation or after three months of treatment, see Figure 1. We also found no such difference for NEDA status at one year. Receiver operator characteristic (ROC) curve analysis found K_i in NAWM to be a good predictor of loss of NEDA status at two years (AUC 0.84, p=0.003). Values of NAWM K_i at six months above a cut-off of 0.136 ml/100/g/min yielded an odds ratio of 11.5 for sub-optimal treatment response at 2 years, with positive and negative predictive values of 73% and 81%, respectively. A one-way repeated measures ANOVA with K_i in NAWM as dependent variable showed a significant treatment effect (p=0.038) only when correcting for baseline methylprednisolone treatment (p=0.0003), days since last relapse (p=0.013), and NEDA status (p=0.045) at two years. Voxel-wise K_i maps are shown in Figure 2, where a widespread general increase at 6 months post-treatment in a subject with loss of NEDA status at 2 years can be appreciated. K_i in NAWM was not correlated with the presence of contrast enhancing lesions, while K_i in NAWM and thalamus was strongly intercorrelated both at baseline (Spearman CC=0.86), 3 months (CC=0.88) and 6 months (CC=0.82), arguing that higher K_i in subjects with sub-optimal treatment response in not the results of spill over of contrast agent from contrast-enhancing lesions into the NAWM. Examples of arterial input functions, tissue curves and patlak plots in two subjects are shown in Figure 4.

Conclusion

Our results suggest that BBB permeability as measured by DCE-MRI reliably predicts sub-optimal treatment response. The observed delay in treatment effect on K_i indicates that DCE-MRI is a surrogate marker of the state of health of the BBB. We find a predictive threshold for disease activity, which is remarkably identical in clinically isolated syndrome, as previously demonstrated² and established RRMS investigated in the present study.

Acknowledgements

We would like to thank radiographers Bente Sonne Møller and Karina Elin Segers at Dept. of Diagnostics, Glostrup Hospital for scanning assistance and the MD consultants Alex Heick, Rikke Jensen, Houry Hassanpour, and Anna Tsakiri at the MS Clinic, Dept. of Neurology, Rigshospitalet, Glostrup for referring patients for the project. We would like to express our gratitude to the patients for participating. Lastly we would like to thank The Research Foundation of the Capital Region of Denmark, Biogen Idec and The Danish Multiple Sclerosis Society, for their financial support, which made this project possible.