MSUS-derived muscle architecture, particularly fiber length, is associated with stroke rehabilitation outcomes and provides incremental predictive value beyond conventional functional assessments.
Key Findings
Results
Favorable rehabilitation outcomes at 6 months were associated with greater tibialis anterior and medial gastrocnemius muscle architectural parameters.
152 stroke patients were included in a single-center retrospective study
105 patients were classified into the favorable outcome group and 47 into the poor outcome group based on predefined composite functional criteria at 6 months after rehabilitation
MSUS parameters assessed included muscle thickness, fiber length, and pennation angle for both the tibialis anterior (TA) and medial gastrocnemius (MG)
Differences in muscle architectural parameters between outcome groups were statistically significant (P < 0.05)
Results
ROC analysis showed moderate predictive ability for individual MSUS fiber length measurements of the tibialis anterior and medial gastrocnemius.
TA fiber length had an AUC of 0.755
MG fiber length had an AUC of 0.652
Both measures demonstrated moderate but not strong standalone predictive performance for rehabilitation outcomes
Results
A combined logistic regression model incorporating multiple MSUS parameters improved prediction of rehabilitation outcomes compared to individual parameters alone.
The combined logistic model achieved an AUC of 0.801
This represented an improvement over the individual parameter AUCs of 0.755 (TA fiber length) and 0.652 (MG fiber length)
The model used multivariable logistic regression incorporating baseline clinical characteristics and MSUS parameters
Results
Adding MSUS parameters to baseline functional scores significantly increased predictive performance in hierarchical regression analysis.
Baseline functional scores alone produced an AUC of 0.721
Adding MSUS parameters to baseline functional scores improved AUC from 0.721 to 0.833
Baseline functional assessments included FMAS, Barthel Index, Brunnstrom stage, MMT, and MAS
Hierarchical regression models were used to evaluate the incremental value of MSUS parameters
Conclusions
Musculoskeletal ultrasound parameters, particularly fiber length, provide incremental predictive value for stroke rehabilitation outcomes beyond conventional functional assessments.
Conventional clinical measures used for comparison included FMAS, Barthel Index, Brunnstrom stage, MMT, and MAS
MSUS-derived muscle architecture was identified as the key contributing parameter, with fiber length specifically highlighted
The study design was single-center and retrospective, with a total sample of 152 patients
ROC analysis, multivariable logistic regression, and hierarchical regression models were all employed to evaluate predictive performance
What This Means
This research suggests that using musculoskeletal ultrasound (MSUS) — a type of imaging that can measure muscle structure — can help predict how well stroke patients will recover over the course of rehabilitation. Specifically, the study measured features of two lower leg muscles (the tibialis anterior at the front of the shin and the medial gastrocnemius, part of the calf) including their thickness, the length of their muscle fibers, and the angle at which those fibers are arranged. Patients who had better muscle fiber lengths in these muscles at baseline tended to have more favorable rehabilitation outcomes six months later.
The study found that while these ultrasound measurements alone were moderately good at predicting outcomes (particularly tibialis anterior fiber length), combining them with standard clinical assessments — such as measures of muscle strength, spasticity, and daily living ability — produced notably better predictions. The combined approach raised the predictive accuracy measure (AUC) from 0.721 using clinical scores alone to 0.833 when ultrasound data was added.
This research suggests that musculoskeletal ultrasound could become a useful tool in stroke rehabilitation settings, giving clinicians additional objective information about muscle health that standard assessments do not capture. By identifying which patients have better or worse muscle architecture early on, rehabilitation teams may be better equipped to tailor treatment plans and set appropriate expectations for recovery. However, as a retrospective single-center study, these findings would benefit from confirmation in larger, prospective, multi-center studies.
Ma C, Liu L, Yang R, Yao P. (2026). Application of musculoskeletal ultrasound in quantitative assessment and rehabilitation of lower leg muscles in stroke patients.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1790982