This study assessed whether quantitative viscoelastic properties of the medial plantar nerve—specifically shear wave speed (Cs) and viscosity coefficient (Vi)—are associated with severity of diabetic peripheral neuropathy (DPN) and can distinguish adjacent clinical severity categories.
Both Cs and Vi were measured using ultrasound-based techniques. The investigators report that Cs and Vi increased progressively across clinical severity groups, supporting a relationship between medial plantar nerve viscoelasticity and DPN severity.
The analysis included 113 patients with type 2 diabetes mellitus. Clinical neuropathy grading used the Toronto Clinical Scoring System (TCSS). Participants were assigned to three groups: non-DPN (n = 33), mild DPN (n = 46), and moderate DPN (n = 34). The source reports no additional demographic or clinical subgroup data in the abstract.
Medial plantar nerve shear wave speed (Cs) was obtained by shear wave elastography, while viscosity imaging provided the viscosity coefficient (Vi). The study combined these measures in logistic regression models to evaluate whether their joint use improves discrimination between adjacent DPN severity categories.
The exact acquisition parameters, probe details, measurement reproducibility, and region-of-interest placement were not reported in the abstract and therefore are not available from the source.
Cs and Vi both increased stepwise from non-DPN to mild DPN to moderate DPN, with statistical significance reported at P < 0.001 for the trend across the three groups. This indicates that higher shear wave speed and higher viscosity are associated with more severe neuropathy in the medial plantar nerve.
No absolute Cs or Vi values for each group are provided in the abstract, and detailed numeric results beyond AUC values for pairwise comparisons were not reported in the source text.
Receiver operating characteristic (ROC) analysis compared Cs, Vi, and a logistic regression–based combined model for discriminating non-DPN from mild DPN. Reported areas under the curve (AUCs) were:
Pairwise comparisons among these AUCs did not reach statistical significance for the non-DPN versus mild DPN comparison, indicating no clear superiority of one metric over another in this adjacent-category discrimination.
For discriminating mild from moderate DPN, reported AUCs were:
Statistical testing showed the combined model provided significantly better discrimination than Cs alone (P = 0.045). However, there was no significant difference between Cs and Vi (P = 0.162) nor between Vi and the combined model (P = 1.000) for this comparison.
The study used DeLong tests to compare AUCs from ROC analyses. Results indicate that viscosity imaging (Vi) had higher AUCs than shear wave speed (Cs) in both pairwise comparisons, and that combining Cs and Vi modestly improved discrimination between mild and moderate DPN relative to Cs alone. The combined model did not significantly outperform Vi alone for mild versus moderate DPN.
No sensitivity, specificity, optimal cutoffs, or net reclassification metrics are reported in the abstract, and the source does not provide further details on model calibration or external validation.
The authors conclude that quantitative medial plantar nerve viscoelastic assessment correlates with DPN severity and that combining shear wave elastography–based Cs and viscosity imaging–based Vi can improve discrimination between mild and moderate DPN compared with Cs alone. They suggest viscoelastic imaging of the medial plantar nerve may complement clinical severity grading.
Limitations evident from the reported information include lack of publicly available datasets (restricted for privacy and ethical reasons) and limited methodological detail in the abstract regarding imaging protocols and measurement reproducibility. The study received ethical approval from the Ethics Committee of the Third Affiliated Clinical Hospital of Changchun University of Chinese Medicine (approval no. CZDSFYLL2026-057-01). The authors declared no competing interests.
Further information such as absolute Cs/Vi group values, technical acquisition parameters, intra- and interobserver reproducibility, and prospective validation would be required to assess clinical implementation and generalizability; these details were not reported in the source abstract.