The research focused on two anatomical territories that define the two faces of MSA. In the parkinsonian variant, known as MSA-P, the pathology concentrates in the basal ganglia, particularly the putamen, a rounded structure deep in the brain that helps initiate and smooth out movement. In the cerebellar variant, MSA-C, degeneration targets the cerebellum and its connections, producing the staggering gait and clumsy coordination that characterize ataxia. Radiologists have long used visual MRI signs to distinguish these patterns: a hyperintense rim around the putamen on diffusion-weighted imaging, bright signal within the lateral putamen on apparent diffusion coefficient maps, and, in the infratentorial compartment, atrophy of the middle cerebellar peduncle and the characteristic hot cross bun sign in the pons.

What makes the new study distinctive is its insistence on structure and reproducibility. Rather than treating these signs as free-form impressions, the researchers scored them systematically. Putaminal rim signs were graded on both diffusion-weighted imaging and susceptibility-weighted imaging, the latter exploiting the magnetic properties of iron deposits that accumulate in the degenerating putamen. Lateral ADC hyperintensity, which reflects increased water diffusion in damaged tissue, was recorded as a binary finding for each side. Infratentorial signs were catalogued separately. Multiple raters assessed the images, and the team measured how well they agreed, finding substantial inter-rater concordance for the visual markers, a result that matters enormously because a sign that different radiologists read differently cannot anchor a diagnosis.

The concordance analysis produced a technically revealing hierarchy. ADC hyperintensity showed only moderate agreement with dichotomized DWI and SWI abnormalities, with Cohen’s kappa values ranging from 0.400 to 0.617, indicating that these markers capture overlapping but not identical tissue changes. By contrast, the DWI and SWI rim scores correlated strongly with each other, with Spearman coefficients between 0.718 and 0.730 and p values below 0.001, suggesting that the rim sign visible on diffusion imaging and the rim sign visible on susceptibility imaging largely reflect the same underlying pathology of the putaminal border. This kind of marker validation is rarely reported in the MSA imaging literature, where signs are often assumed rather than tested.

The distribution of signs across subtypes confirmed expectations. Putaminal abnormalities were more frequent in patients with MSA-P, while infratentorial signs dominated in MSA-C, reinforcing the anatomical logic that separates the two variants. But the study’s central question was more ambitious: does the side of the brain showing the abnormal signal predict the side and severity of the motor deficit? Because the basal ganglia control the contralateral body, an abnormality in the left putamen should, in principle, produce worse parkinsonian signs in the right limbs. Testing this requires side-specific motor scores, which were available for 51 of the 94 patients, and heel-knee-shin ataxia scores for 49 patients, a limitation the authors acknowledge candidly.

The initial, unadjusted analyses hinted at a lateralized relationship: left-sided ADC hyperintensity appeared associated with higher contralateral motor scores. But when the researchers built models adjusted for age and MSA subtype, analyzing each side separately, the associations dissolved. Separate side-specific putaminal models did not show significant relationships with the contralateral total lateral motor score, and infratentorial sign models failed to reach significance against the heel-knee-shin score. This pattern is a familiar lesson in clinical statistics: crude associations can reflect confounding by subtype, age, or disease severity, and only adjusted models reveal whether a signal is genuine.

The breakthrough came from a different modeling strategy. Instead of analyzing each hemisphere in isolation, the team pooled both sides using generalized estimating equations, a statistical framework designed for correlated repeated measures, here treating the left and right hemispheres of each patient as paired observations. These side-combined GEE models, adjusted for age, MSA subtype, and imaging side, delivered consistent results: ADC hyperintensity was associated with higher contralateral total lateral motor scores with a p value of 0.004, and higher DWI and SWI rim scores were similarly associated with p values of 0.017 and 0.015 respectively. In other words, a putamen that looks abnormal on MRI tends to belong to a body whose opposite limbs move more poorly, exactly what the neuroanatomy predicts.

The authors are careful about what these findings do and do not establish. The GEE results suggest genuine putaminal imaging-motor associations, but the study was cross-sectional, capturing a single moment in each patient’s disease course, and the incomplete side-specific clinical assessments limit how broadly the conclusions generalize. The diagnostic value of these signs, the team emphasizes, requires evaluation in comparator cohorts that include patients with Parkinson’s disease, progressive supranuclear palsy, and other parkinsonian mimics, since a sign that correlates with motor severity within MSA may or may not distinguish MSA from its imitators. That distinction, between phenotypic characterization and differential diagnosis, is the crux of how such markers will ultimately be used in clinics.

For the field, the study offers something quietly valuable: evidence that structured visual MRI assessment provides reproducible information for characterizing MSA patients, without requiring advanced post-processing or quantitative pipelines that many centers cannot deploy. The signs that neuroradiologists have read by eye for decades, the putaminal rim, the bright lateral ADC signal, the atrophic cerebellar peduncle, are not arbitrary impressions but measurable correlates of the clinical picture. As MSA research pushes toward earlier diagnosis and disease-modifying trials, the ability to link what the scanner sees to what the patient experiences, side by side and hemisphere by hemisphere, becomes a foundation for stratifying patients and tracking treatment effects. The work from Beijing adds rigorous statistical footing to that effort, one kappa and one p value at a time.

Subject of Research: Association of visually assessed putaminal and infratentorial MRI signs with motor symptoms in multiple system atrophy

Article Title: Association of visually assessed putaminal and infratentorial MRI signs with motor symptoms in multiple system atrophy: a cross-sectional study

Article References: Li, K., Zhao, Y., Zheng, Y., Luan, M., Tang, H., He, C., Liu, J., Li, F., Wei, L., Chen, J., Sun, W., Deng, J., Wang, Z., Zhu, Y., & Sun, Y. (2026). Association of visually assessed putaminal and infratentorial MRI signs with motor symptoms in multiple system atrophy: a cross-sectional study. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02958-0

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02958-0

Keywords: multiple system atrophy, MRI, putamen, apparent diffusion coefficient, diffusion-weighted imaging, susceptibility-weighted imaging, parkinsonism, cerebellar ataxia, neuroradiology, generalized estimating equations, MSA-P, MSA-C