Investigating the correlation between temporalis, masseter and cervical muscle mass on a single cross-sectional image
Year: 2026
Type: article
Study design: Observational Study
Abstract: Background: Low skeletal muscle mass (SMM) is an established prognostic factor in oncology and is conventionally quantified at the third lumbar vertebra. In brain tumour patients, abdominal imaging is seldom available, and craniofacial measurements from routine brain magnetic resonance imaging (MRI), temporalis muscle thickness (TMT) and masseter cross-sectional area (MCSA), have been proposed as alternatives. Whether these measurements can substitute for one another, or for cervical [skeletal muscle area at the third vertebra (C3SMA)] and lumbar muscle mass, is unknown. We investigated the correlation and classification agreement between these measurements. Methods: A retrospective cross-sectional study was conducted on 247 patients who underwent awake craniotomy at Amsterdam University Medical Center between 2003 and 2020. TMT, MCSA and C3SMA were measured on pre-operative three-dimensional (3D) T1-weighted MRI, and L3 skeletal muscle area (L3SMA) and index (L3SMI) were estimated from C3SMA using a previously established method. TMT was available in 247 patients, MCSA in 239 and C3SMA in 136. Associations were assessed using Spearman correlation coefficients, crude and adjusted for body mass index (BMI) and, separately, for age and sex. Low SMM was classified using established cut-off values for TMT (≤6.3 mm in men, ≤5.2 mm in women) and L3SMI (<43.2 cm 2 /m 2 ) and, as no validated masseter threshold exists, the sex-specific lowest quartile for MCSA. Agreement between classifications was assessed using Cohen kappa and interobserver reliability using intraclass correlation coefficients (ICC). Results: All pairwise correlations were low: TMT and MCSA r=0.39 (P<0.05, n=239), MCSA and C3SMA r=0.30 (P<0.05, n=136), TMT and estimated L3SMA r=0.24 (P<0.05, n=136) and TMT and C3SMA r=0.20 (P<0.05, n=136). Adjustment for BMI altered the coefficients minimally (TMT and MCSA r=0.37, P<0.001), whereas adjustment for age and sex rendered the association between MCSA and C3SMA non-significant (r=0.11, P=0.20). Among the 136 patients with all three measurements, agreement in classifying low SMM was poor using the thresholds applied (Cohen kappa 0.11 for TMT versus MCSA, 0.07 for MCSA versus L3SMI and 0.01 for TMT versus L3SMI), and only 2 patients (1.5%) were concurrently classified as having low SMM by all three measurements. Interobserver reliability was excellent [ICC 0.919, 95% confidence interval (CI): 0.790–0.969 for TMT; ICC 0.987, 95% CI: 0.967–0.995 for MCSA]. Conclusions: Craniofacial, cervical and estimated lumbar muscle measurements showed low correlations with one another and poor agreement in classifying low SMM using the thresholds applied in this study. As L3SMI was estimated rather than directly measured, non-interchangeability cannot be definitively established, but these findings provide little support for using craniofacial measurements as substitutes for one another or for cervical and lumbar muscle mass.
Language: English
Subfield: Physiology (medical)
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7223245296