Assessing skeletal muscle mass based on the cross-sectional area of muscles at the 12th thoracic vertebra level on computed tomography in patients with oral squamous cell carcinoma

Oral Oncol. 2021 Feb:113:105126. doi: 10.1016/j.oraloncology.2020.105126. Epub 2020 Dec 31.

Abstract

Objectives: This study aimed to create a formula to estimate the third lumbar vertebra (L3)1 level skeletal muscle cross-sectional area (CSA), known as a standard value to evaluate skeletal muscle mass on computed tomography (CT), using the twelfth thoracic vertebra (Th12) level skeletal muscle CSA on chest CT.

Materials and methods: This retrospective observational study included patients aged 40 + years with a diagnosis of oral squamous cell carcinoma (n = 164). Skeletal muscle CSA on CT images was measured using the Th12 and the L3 levels of pretreatment CT scans. The predictive formula was created based on the five-fold cross-validation method with a linear regression model. Correlations between the predicted L3-level CSA and the actual L3-level CSA were evaluated using r and Intraclass Correlation Coefficients (ICC).

Results: The predictive formula for L3-level CSA from Th12-level CSA was: CSA at L3 (cm2) = 14.143 + 0.779 * CSA at Th12 (cm2) - 0.212 * Age (y) + 0.502 * Weight (kg) + 13.763 * Sex. Correlations between the predicted and measured L3-level CSA were r = 0.915 [0.886-0.937] and ICC = 0.911 [0.881-0.934].

Conclusion: We developed a formula for predicting skeletal muscle mass from the Th12-level CT slice. The predicted L3-level CSA correlated with the measured L3-level CSA.

Keywords: Estimation formula; Head and neck tumor; Oral cancer; Sarcopenia; Skeletal muscle mass.

MeSH terms

  • Aged
  • Carcinoma, Squamous Cell / diagnostic imaging*
  • Carcinoma, Squamous Cell / pathology
  • Female
  • Humans
  • Lumbar Vertebrae / physiopathology*
  • Male
  • Mouth Neoplasms / diagnostic imaging*
  • Mouth Neoplasms / pathology
  • Muscle, Skeletal / physiopathology*
  • Retrospective Studies
  • Tomography, X-Ray Computed / methods