Masticatory muscle index for indicating skeletal muscle mass in patients with head and neck cancer.

Background: A typical assessment for sarcopenia involves the use of abdominal computed tomography (CT) for calculating the skeletal muscle index (SMI) at the level of the third lumbar vertebra (L3). However, abdominal CT is not regularly performed on patients with head and neck cancer (HNC). We investigated whether masticatory SMI (M-SMI) measurements based on head and neck CT scans can be used to conduct sarcopenia assessments by evaluating whether M-SMI is correlated with L3-SMI.

Methods: Abdominal and head and neck CT images of patients with trauma (n = 50) and HNC (n = 52) were analyzed retrospectively. Both manual delineation and threshold selection methods were used to measure cross-sectional areas of masticatory muscles and those of muscles at the L3 level on CT images. Muscle cross-sectional areas were normalized to height squared to calculate SMI, and a multivariate linear regression model was established to evaluate the correlation between the M-SMI and L3-SMI. Receiver operating characteristic curve analysis was used to assess the ability of the M-SMI to identify sarcopenia, and Cox logistic regression was used to identify predictors of sarcopenia.

Results: Patients with HNC had significantly lower M-SMI and L3-SMI than did patients with trauma (p = 0.011 and 0.03, respectively). M-SMI and L3-SMI were strongly correlated (r = 0.901, p < 0.001); in the multivariate model that included sex, the correlation was stronger (r = 0.913, p < 0.001). The associations of sarcopenia with a lower M-SMI (p < 0.001), male sex (p = 0.028), and advanced age (p = 0.011) were significant, and multivariate logistic analysis demonstrated that an M-SMI of <5.5 was an independent predictor of sarcopenia (hazard ratio = 5.37, p < 0.001).

Conclusions: M-SMI assessment in routine head and neck CT scans is feasible and can be an alternative for detecting sarcopenia in patients with HNC.