To evaluate the relationship between measures of muscle assessed by computed tomography (CT) and dual-energy x-ray absorptiometry (DXA) and to determine the agreement between techniques for categorising low muscle mass in patients with Malignant Pleural Mesothelioma (MPM).
This retrospective study used data from two studies of patients with MPM. Inclusion criteria were: availability of CT scan and DXA ≤14 days apart. Exclusion criteria were: CT scan of inadequate quality or third lumbar vertebrae not visible. Cross-sectional images at the first and third lumbar vertebrae were identified, and muscle area analysed using Slice-O-Matic software v5.0 (Tomovision, Montreal, QC, Canada). Low muscle mass was defined via CT (skeletal muscle index <55.4 cm2/m2 for men and <38.9 cm2/m2 for women) and DXA (appendicular lean mass <7.26 kg/m2 for men and <5.45 kg/m2 for women). Spearman’s correlations were used to evaluate the relationship between cross-sectional muscle area and whole-body lean mass. Kappa (ĸ) coefficient was calculated to assess the agreement between low muscle mass criteria.
Of 54 potentially eligible CT scans, 49 (91%) were included [reasons for exclusion: inadequate quality (n=2); third lumbar vertebrae not visible (n=3)]. CT cross-sectional muscle area at the first and third lumbar vertebrae was strongly positively correlated with whole-body lean mass on DXA (rho=0.803, rho=0.796, p<0.001, respectively). Agreement between low muscle mass criteria was 63.3% (n=31, ĸ=0.211). The CT criteria had high sensitivity (93%) but low specificity (27%) for categorising low muscle mass when compared with DXA.
Our results indicate that in patients with MPM, cross-sectional muscle area on CT images is strongly correlated with whole-body lean mass on DXA. However, there is poor agreement between criteria for categorising low muscle mass. This novel technique of analysing existing clinical CT scans can be used for future research on the relationship between body composition and patient outcomes in MPM.