151751-Najiba-Chargi

35 Diagnostics: validation of skeletal muscle mass measurement Figure 1. Delineation of skeletal muscle tissue on transversal CT imaging at the level of L3 (A) and at the level of C3 (B). A Hounsfield Unit window of -29 to +150 was used to accentuate skeletal muscle tissue. CROSS-SECTIONAL SKELETAL MUSCLE AREA AT THE LEVEL OF L3 As well as the actually measured SMA at the level of L3, the SMA at the level of L3 was also estimated from the SMA at the level of C3 using the prediction formula as described by Swartz et al, see formula 1. 7 Formula 1: SMA at L3 (cm 2 ) = 27.304 + 1.363 * CSA at C3 (cm 2 ) – 0.671*Age (years) + 0.640 * Weight (kg) + 26.442*Gender (Gender=1 for female and 2 for male) The lumbar SMI was then calculated using the formula published by Prado et al, see Formula 2. 5 Formula 2: LSMI (cm 2 /m 2 ) = SMA at L3/(height*height) Low SMM was defined as a LSMI ≤43.2cm 2 /m 2 , as previously published by Wendrich et al. 8 STATISTICAL ANALYSIS All statistical analyses were performed using the IBM SPSS Statistics version 25.0 software package (Chicago, Illinois, USA). There were nomissing data. A test for normality (Shapiro-Wilk test) was performed to assess whether continuous variables were normally distributed. For table 1 continuous data are represented as mean ± standard deviation (SD) if normally dis - tributed, and median ± range if skewed. Categorical data are represented as a number and percentage of total. The student’s t -test, one-way ANOVA, Mann-Whitney U test were used where appropriate. Depending on normality of variables, Pearson or Spearman Rank correla - tion coefficients were calculated to assess correlation between SMA at the level of C3, at the level of L3 and predicted SMA at the level of L3. To assess the agreement between measure- ments, we calculated intraclass correlation coefficients (ICCs) using a two-way mixed single 2