CT-defined sarcopenia, or low skeletal muscle mass, assessed using the cross-sectional area (CSA) at the third lumbar vertebra (L3), is the gold standard in body composition analysis and an independent prognostic indicator in head and neck cancer (HNC). Body mass index (BMI) is often used as a nutritional marker, however this measure does not indicate proportions of lean mass, and sarcopenic obesity (obesity with depleted muscle) can go undiagnosed. This study compared two methods of sarcopenia assessment in patients with HNC who are overweight or obese.
All patients presenting to the head and neck clinic until December 2020 with newly diagnosed HNC of the larynx, hypopharynx, nasopharynx, oropharynx or oral cavity, with a BMI>25 who had a diagnostic positron emission tomography-computed tomography (PET-CT) scan were included. CSA of muscle was measured at L3 and the third cervical vertebra (C3) and converted to an estimated L3 value (described by Swartz et al. 2016). Agreement between methods was assessed by the Bland-Altman method. Sarcopenia was determined based on pre-defined, sex-specific thresholds for skeletal muscle index and compared in each group.
Scans of 59 patients were analysed, of which 26(44%) had a BMI ≥30 (92% male, 76% oropharynx, median age 59yrs). Good correlation was observed between L3 and estimated C3 CSA (r=0.78, p<0.001). Mean difference (bias) =12.0cm2, (SD=20.4, 95%CI 6.7-17.3) with limits of agreement exceeding clinical acceptability (-27.9-40.9cm2), indicating poor agreement between methods. Sarcopenia was diagnosed in 19% of the L3 group and 39% in the C3 group (sensitivity 72.7%, specificity 68.8%). There was weak agreement in sarcopenia diagnosis (ƙ=0.292, 95%CI 0.1-0.5).
The level of agreement between the C3 estimate and actual L3 measures of skeletal muscle in this cohort is weak. The subsequent diagnosis of sarcopenia may be substantially inaccurate and requires further validation in larger populations.