Myofascial pain syndrome (MPS) diagnosis is currently a source of contention amongst clinicians. Physicians, chiropractors, and registered massage therapists do not agree on the criteria they deem diagnostic of MPS. Ultrasound imaging is widely used in medicine to qualitatively identify anatomical structures and lesions. Quantitative ultrasounds methods such as texture features have been used to characterize normal and pathological tissue in muscular disorders. We propose that quantitative ultrasound imaging techniques can be used to differentiate muscle with myofascial pain diagnosis from healthy controls. Therefore, we assessed whether texture features could differentiate upper trapezius muscle in patients with MPS relative to healthy participants.

We collected B-mode ultrasound images of the upper trapezius muscle in 15 healthy participants and 17 patients with MPS. The following texture features were extracted from the images: blob area, blob count, and 10 local binary patterns (LBP) (Fig. 1). A principal components analysis (PCA) was performed to reduce the features to those that accounted for the most variability in the images (Table 1). A MANOVA was then performed to determine whether healthy or MPS group membership could be differentiated by the reduced features (Table 2).

The PCA identified two components that accounted for 93% of the variability. Features with the highest loading factors included: LBP2, LBP6, LBP10, blob area, and blob count. All features but blob area could statistically differentiate healthy from myofascial pain groups (P<0.001).

Texture features are capable of differentiating muscle tissue in patients with MPS relative to healthy individuals. These findings may be used to develop an accurate diagnostic tool for clinicians to diagnose patients with MPS.