Free-water MRI

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

David Vaillancourt's group at the University of Florida (earlier at the University of Illinois at Chicago), with Edward Ofori, Roxana Burciu, Ofer Pasternak and colleagues.

What happened

Diffusion MRI tracks how water moves in tissue. When nerve cells and fibres die, extra water sits between them. A two-compartment ("bi-tensor") model separates that free water from tissue.

• 2009. Vaillancourt et al. used a simpler high-resolution diffusion method (fractional anisotropy) in the substantia nigra of 14 untreated early patients and 14 controls; it separated them, strongest in the caudal nigra, with 100% sensitivity and specificity in that region (Neurology 72:1378–1384). This was not yet free water. The paper has a published erratum (Neurology 2009;72:2059, DOI 10.1212/01.wnl.0000344311.87440.f5); its content could not be read.
• 2015. Ofori et al. found raised free water in the posterior nigra in patients versus controls, at one site and across multiple sites (Neurobiol Aging 36:1097–1104). A second paper followed 25 patients and 19 controls for a year: free water rose in patients only (P=0.006), and baseline free water predicted the 1-year change in bradykinesia (r=0.74) (Brain 138:2322–2331).
• 2017. Burciu et al. used PPMI data (landmark 24): 103 new patients and 49 controls at 1 year, and 46 patients to 4 years. Free water rose in patients, not controls, and kept rising; 1- and 2-year rises predicted worse Hoehn and Yahr stage at 4 years and matched falling putamen dopamine-transporter binding (Brain 140:2183–2192). The same year a 425-person multisite paper (107 controls, 184 PD, 63 MSA, 71 PSP) showed free water was higher in PSP than in PD and MSA (Mov Disord 32:1457–1464).
• Neuromelanin MRI. A separate approach images neuromelanin, the pigment in dopamine neurons. Sasaki et al. (2006, Neuroreport) showed reduced signal in the locus ceruleus and substantia nigra in Parkinson's. That paper is cross-sectional, not a progression study.

Why it moved us toward a cure

A trial of a slowing drug takes years if it relies on symptom scores. An imaging marker that changes within a year, and that tracks long-term decline, could shorten trials. Free water was the main endpoint of the University of Florida rasagiline imaging trial (Arpin et al., Mov Disord 2022, online Nov 2021; NCT02789020) described in landmark 62.

Outcome

Open. That trial (90 and 75 scans analysed on two scan settings) found no group difference (P = 0.346 and 0.228), which does not show whether the marker is insensitive or the drug inactive. The main papers come from one group, with only the 2017 multisite study and PPMI data reaching beyond Florida. A 2018 Florida study of later-stage patients (Guttuso et al.) found increases only in the anterior nigra, not the posterior. No regulatory qualification of free water as a trial endpoint was found, and its link to cell loss in the nigra is indirect. No correction or retraction was found for the 2015 and 2017 papers in Europe PMC.

Sources

• Vaillancourt et al., Neurology 2009, DOI 10.1212/01.wnl.0000340982.01727.6e
• Ofori et al., Neurobiol Aging 2015, DOI 10.1016/j.neurobiolaging.2014.10.029 · Ofori et al., Brain 2015, DOI 10.1093/brain/awv136
• Burciu et al., Brain 2017, DOI 10.1093/brain/awx146 · Ofori et al., Mov Disord 2017, DOI 10.1002/mds.27100
• Sasaki et al., Neuroreport 2006, DOI 10.1097/01.wnr.0000227984.84927.a7
• Arpin et al., Mov Disord 2022;37:325–333, DOI 10.1002/mds.28838 · NCT02789020
• Not yet verified: a neuromelanin MRI paper showing change over time in Parkinson's (not found); whether any trial other than the rasagiline study has used free water as its primary endpoint; the 2018 Guttuso study (seen only in a search snippet); any regulatory qualification status.